| *** dims_ has quit IRC | 00:00 | |
| arosen | I can submit a pull request to remove those lines. | 00:00 |
|---|---|---|
| asselin | arosen, daya_k I've solved quite a few issues in a fork of jaypipes's repo & added some features (nodepool, log server) https://github.com/rasselin/os-ext-testing/ | 00:00 |
| *** wenlock has quit IRC | 00:01 | |
| daya_k | asselin: thanks, i'll check it out. | 00:01 |
| *** sarob_ has joined #openstack-infra | 00:01 | |
| asselin | we were thinking moving it to https://github.com/stackforge | 00:01 |
| daya_k | fungi: its still dying if i clean out git dir, kill zuul-merger and restart as a service. works ok with -d, will look out for the exception,its intermittent | 00:02 |
| *** mmaglana has quit IRC | 00:02 | |
| fungi | asselin: so it *looks* like maybe we've grown system dependencies on libxml2-dev and libxslt1-dev which is what's tanking the pip install. we probably need to conditionally ensure those are present in the zuul:init class (pip can't do anything about them since they're c lib headers) | 00:02 |
| daya_k | asselin : i'm using jaypipes repo, which pulls from https://git.openstack.org/cgit/openstack-infra/config/plain/install_puppet.sh | 00:03 |
| asselin | daya_k, yes, my fork didn't change that part | 00:04 |
| fungi | asselin: i believe this is recently introduced. looking at the log lxml is being pulled in by python-keystoneclient which in turn is for swift object support we just added | 00:04 |
| arosen | asselin: if i change layout.yaml what all do i need to restart just all the zuul stuff i guess? | 00:04 |
| asselin | arosen, you should rerun the install_master script so the changes is updated in /etc/zuul/ the puppet script will signal to zuul to reload | 00:05 |
| *** sarob_ has quit IRC | 00:05 | |
| sweston | fungi, asselin, daya_k: fungi's assertion is correct, I believe. I can verify this tonight in a fresh vm | 00:06 |
| *** jaypipes has quit IRC | 00:06 | |
| arosen | asselin: k | 00:06 |
| *** crc32 has quit IRC | 00:06 | |
| *** markmcclain has quit IRC | 00:06 | |
| arosen | asselin: hrm it's saying: Could not parse configuration file: Certificate names must be lower case; see #1168 when I run: sudo bash install_master.sh | 00:07 |
| daya_k | arosen : also, for debugging, sweston gave this seq could help http://paste.openstack.org/show/89060/ | 00:07 |
| asselin | arosen, never seen that error. something unrelated to the layout.yaml change | 00:09 |
| fungi | clarkb: jhesketh: do you recall anything we might have done on our zuul servers to avoid the pip install errors bubbling up from keystoneclient pulling in lxml? http://paste.openstack.org/show/89055/http://paste.openstack.org/show/89055/ | 00:09 |
| fungi | gah, paste button is bouncing again... http://paste.openstack.org/show/89055/ | 00:10 |
| jhesketh | fungi: no idea in regards to our zuul servers, but in a past life I've had trouble with compiling lxml and the solution was just to install it from the os packages | 00:11 |
| fungi | jhesketh: yeah, i'm not seeing that we've done that either | 00:11 |
| fungi | ohhhhh | 00:12 |
| fungi | no, nevermind. thought i saw something | 00:12 |
| *** vermarah_ has quit IRC | 00:12 | |
| *** banix has joined #openstack-infra | 00:13 | |
| fungi | we do have libxml2 installed from system packages, but no pip install of lxml nor system package of python-lxml | 00:13 |
| arosen | asselin: do you know why recheck no bug or recheck congress isn't working here: http://paste.openstack.org/show/89061/ | 00:14 |
| asselin | arosen, yes, the regex is not correct | 00:14 |
| jhesketh | arosen: oh, you're around :-). I'm still waiting on a response to https://review.openstack.org/#/c/77347/5 if you have time please | 00:15 |
| jhesketh | fungi: completely unrelated, I'm having trouble with the latest global requirements https://bugs.launchpad.net/nova/+bug/1350107 | 00:15 |
| fungi | arosen: jaypipes wrote that article back when we were still running gerrit 2.4, and the comment format has changed in the event stream since then with gerrit 2.8 | 00:15 |
| uvirtbot | Launchpad bug 1350107 in nova "Could not find a version that satisfies the requirement oslo.config>=1.4.0.0a3" [Undecided,In progress] | 00:15 |
| asselin | arosen, try: comment: (?i)^(Patch Set [0-9]+:)?( [\w\\+-]*)*(\n\n)?\s*recheck(( (?:bug|lp)[\s#:]*(\d+))|( MyVendor))\s*$ | 00:15 |
| jhesketh | fungi: basically oslo.config 1.4.0.0a3 isn't on pypi and I'm not sure how it's meant to be put there etc | 00:16 |
| *** arnaud has joined #openstack-infra | 00:16 | |
| *** arnaud__ has joined #openstack-infra | 00:16 | |
| fungi | jhesketh: the bug is half wrong and half misguided ;) | 00:17 |
| asselin | fungi, I wanted to bring it up in the infra meeting today. About migrating jaypipe's repo to stackforge where it's easier for everyone to work and fix issues. | 00:17 |
| jhesketh | fungi: I thought it might be | 00:17 |
| fungi | jhesketh: jhesketh it *is* on pypi.python.org | 00:17 |
| fungi | jhesketh: and as for pypi.openstack.org you should be installing from /simple now not /openstack | 00:17 |
| arosen | asselin: not comment_filter ? | 00:17 |
| asselin | arosen, both work...comment_filter is the old way | 00:18 |
| lifeless | mordred: reminder - 94108 needs review | 00:18 |
| jhesketh | fungi: right, but only the wheels are on there? | 00:18 |
| fungi | asselin: i encouraged jaypipes and others to update and start managing that walkthrough in http://ci.openstack.org/third_party.html | 00:18 |
| *** unicell has joined #openstack-infra | 00:19 | |
| fungi | jhesketh: we only upload wheels for prereleases because pip<1.4 doesn't know not to upgrade to them automatically | 00:19 |
| fungi | jhesketh: but pip<1.4 also doesn't know about wheels at all | 00:19 |
| fungi | jhesketh: this way we get the prerelease only when explicitly requested with newer pip versions and not at all with older pip versions | 00:20 |
| jhesketh | fungi: so the turbo-hipster workers have pip 1.5.6 but fail to pull in the required oslo.config | 00:20 |
| *** gokrokve has quit IRC | 00:21 | |
| fungi | jhesketh: are they going to http://python.pypi.org/simple? | 00:21 |
| *** gokrokve has joined #openstack-infra | 00:21 | |
| fungi | hurr, pypi.python.org | 00:21 |
| *** banix has quit IRC | 00:21 | |
| daya_k | fungi: actually i had a suggestion, some pre-built vm image templates could be hosted with all of this setup, so 3rd party ci could kick off | 00:22 |
| jhesketh | fungi: https://pypi.python.org/simple/oslo.config/ | 00:22 |
| jhesketh | fungi: https://ssl.rcbops.com/turbo_hipster/result_server.fcgi?q=turbo_hipster/results/60/109660/7/check/real-db-upgrade_nova_mysql_devstack_131007:th-mysql/be6e5c82a6f843f289954ee49dcb8a52/131007_devstack_export.log | 00:22 |
| *** yamahata_ has joined #openstack-infra | 00:23 | |
| *** gangil has quit IRC | 00:23 | |
| asselin | fungi, sorry a bit confused. I don't see anything related to stackforge in that link...other than the link to this http://ci.openstack.org/stackforge.html? | 00:23 |
| *** julim has quit IRC | 00:23 | |
| r1chardj0n3s | fungi: that pip<1.4 hack is only going to last so long. Perhaps it might be worth looking into a devpi to allow staged release of alpha/beta packages? | 00:24 |
| *** pcm_ has quit IRC | 00:24 | |
| *** pcm_ has joined #openstack-infra | 00:24 | |
| clarkb | I dont think we need to. pip does the correct thing with alpha versions | 00:25 |
| *** gokrokve has quit IRC | 00:25 | |
| r1chardj0n3s | clarkb: yep, cool | 00:25 |
| clarkb | <1.4 didnt hence wheels | 00:25 |
| fungi | clarkb: r1chardj0n3s is referring to the europython discussions around possibly removing wheel-only uploads from teh cheeseshop (and the subsequent discussion on distutils-sig) | 00:26 |
| clarkb | so wheels are important in that way | 00:26 |
| fungi | presumably | 00:26 |
| *** Alexandra_ has joined #openstack-infra | 00:26 | |
| *** sarob_ has joined #openstack-infra | 00:26 | |
| r1chardj0n3s | fungi: yeah, though I think maybe that has less traction that I'd thought | 00:26 |
| clarkb | meh you dont have to provide src at download time usually | 00:26 |
| clarkb | even gpl is fine that way | 00:26 |
| r1chardj0n3s | pip really needs to get a damn regular self-update check built into it ASAP | 00:27 |
| fungi | r1chardj0n3s: i don't think the cheeseshop can effectively be the free software license police. you can already upload wheels with no tarball. anyway, conversation for another time | 00:27 |
| r1chardj0n3s | clarkb: mostly it's for the downstream packagers | 00:27 |
| arosen | is there anyway to check if the layout.yaml has an errors? I made a few changes trying to fix the comment and now it seems nothing is working. I haven't found any errors in the logs either :/ | 00:27 |
| fungi | er, upload eggs with no tarball | 00:27 |
| fungi | could already before there were wheels | 00:27 |
| r1chardj0n3s | fungi: yeah, that's a strong argument. as I mentioned, the idea was mostly to help downstream packagers | 00:27 |
| clarkb | arosen there is a voluptuous check you can run | 00:29 |
| clarkb | it makes sure the layour is syntactically sound | 00:29 |
| jhesketh | fungi, r1chardj0n3s: so any thoughts to why I can't pip install the olso.config package? | 00:29 |
| fungi | asselin: i was actually talking about moving the walkthrough into that document, to update it to be more current and useful | 00:29 |
| r1chardj0n3s | jhesketh: is oslo.config already installed? what command are you running to try to install it? | 00:29 |
| *** garyh has joined #openstack-infra | 00:30 | |
| asselin | fungi, and by walkthrough you mean his blogs? | 00:30 |
| jhesketh | r1chardj0n3s: it shouldn't be, but the error is about the version anyway | 00:30 |
| *** sarob_ has quit IRC | 00:30 | |
| jhesketh | r1chardj0n3s: pip install -q -r $requires | 00:30 |
| jhesketh | where $requres is requirements.txt | 00:30 |
| arosen | alright well i think i'm going to break for dinner. Thanks guys for your help here. Hopefully I'll make some more progress tmw :) | 00:31 |
| r1chardj0n3s | ok, and what does oslo.config look like in the requirements.txt? | 00:31 |
| r1chardj0n3s | BTW "pip install 'oslo.config>=1.4.0.0a3'" works for me | 00:31 |
| fungi | jhesketh: your log says you're using http://pypi.openstack.org/openstack rather than http://pypi.openstack.org/simple | 00:31 |
| *** yamahata_ has quit IRC | 00:31 | |
| r1chardj0n3s | but only because pip is being forced to consider that alpha version | 00:31 |
| *** yamahata_ has joined #openstack-infra | 00:31 | |
| fungi | jhesketh: also you can just use https://pypi.python.org/simple now anyway | 00:31 |
| fungi | asselin: yes, blogs | 00:32 |
| asselin | fungi, ok, yes that makes sense. | 00:32 |
| fungi | jhesketh: pypi.o.o/openstack was our old restricted mirror, which is now abandoned and no longer updated. we might should just delete its contents to make the broken more obvious at this point | 00:33 |
| *** bdpayne has quit IRC | 00:33 | |
| fungi | jhesketh: pypi.o.o/simple is the new full pypi mirror we're using for upstream jobs | 00:33 |
| jhesketh | fungi: I've got a fix up for going ti simple | 00:33 |
| *** Alexandra_ is now known as a | 00:34 | |
| jhesketh | fungi: but either way pypi.o.o/simple is also configured | 00:34 |
| *** a is now known as Guest88745 | 00:34 | |
| jhesketh | fungi: r1chardj0n3: http://paste.openstack.org/show/89062/ | 00:34 |
| asselin | I think it would be helpful to move his project to stackforge. it's out of date, and keeping it up to date is easier to do with o.o's infrastructure. It would also make it easier to maintain this link: http://ci.openstack.org/third_party.html | 00:35 |
| fungi | jhesketh: the other possibility is that you're using virtualenv earlier than 1.11, and so it's using pip 1.4.x in a virtualenv which ignores wheels unless told to --use-wheel | 00:35 |
| jhesketh | fungi: hmm, good thought... let me look | 00:35 |
| jhesketh | fungi: well called, it's 1.10.1 | 00:36 |
| *** Guest88745 is now known as alexandra | 00:36 | |
| fungi | jhesketh: yeah, we install 1.11.6 on our test nodes currentl;y | 00:36 |
| jhesketh | okay, so I need to upgrade that and hopefully that'll fix things | 00:36 |
| jhesketh | fungi: how are you getting the newer version? | 00:36 |
| *** jgrimm has joined #openstack-infra | 00:36 | |
| fungi | jhesketh: pip i believe... checking now | 00:36 |
| jhesketh | fungi: thanks, if it's in the config repo let me know where it is :-) | 00:37 |
| fungi | looks like maybe the openstack_project::base class | 00:37 |
| *** zz_dimtruck is now known as dimtruck | 00:37 | |
| jhesketh | fungi: thanks :-) | 00:38 |
| *** armax has quit IRC | 00:38 | |
| fungi | jhesketh: also devstack does a 'pip install -U virtualenv' in stack.sh (via the pip_install function) | 00:39 |
| fungi | jhesketh: but i assume t-h migrations jobs don't use devstack at all | 00:39 |
| jhesketh | nope, but that's interesting | 00:39 |
| jhesketh | we could do our own -U in th, but I think I'll put it into our puppet | 00:39 |
| fungi | so yeah, it looks like maybe we use 1.11.4 on our bare nodes and the latest (which is currently 1.11.6) on devstack nodes | 00:40 |
| fungi | probably for no particular reason | 00:40 |
| *** daya_k has quit IRC | 00:42 | |
| *** arnaud__ has quit IRC | 00:42 | |
| *** arnaud has quit IRC | 00:42 | |
| jhesketh | thanks for your help fungi :-) | 00:43 |
| *** banix has joined #openstack-infra | 00:44 | |
| fungi | jhesketh: yw | 00:44 |
| *** sarob_ has joined #openstack-infra | 00:46 | |
| *** banix has quit IRC | 00:46 | |
| *** sarob_ has quit IRC | 00:50 | |
| *** banix has joined #openstack-infra | 00:50 | |
| *** amcrn has quit IRC | 00:52 | |
| *** dhellmann is now known as dhellmann_ | 00:53 | |
| *** jp_at_hp has quit IRC | 00:53 | |
| *** shashankhegde has joined #openstack-infra | 01:01 | |
| *** shashankhegde has quit IRC | 01:01 | |
| reed | how can I check the load of lists.openstack.org? it seems a bit too slow lately (the web archives) | 01:06 |
| clarkb | reed http://cacti.openstack.org | 01:09 |
| reed | clarkb, thanks | 01:09 |
| reed | clarkb, I think slowness is on my side | 01:10 |
| *** arnaud has joined #openstack-infra | 01:10 | |
| *** griff is now known as jgriffith | 01:11 | |
| *** rainya has quit IRC | 01:11 | |
| fungi | doesn't look loaded at all.. nearly idle | 01:11 |
| reed | yep | 01:12 |
| reed | oh well, I was reminded of cacti again today #success | 01:12 |
| *** homeless has quit IRC | 01:13 | |
| *** camunoz has quit IRC | 01:13 | |
| StevenK | fungi, clarkb: Can you two please poke at https://review.openstack.org/#/c/108817/ ? | 01:14 |
| greghaynes | yesplz | 01:14 |
| *** rwsu has quit IRC | 01:18 | |
| *** yamahata_ has quit IRC | 01:26 | |
| *** camunoz has joined #openstack-infra | 01:26 | |
| *** marcoemorais has quit IRC | 01:27 | |
| *** yamahata_ has joined #openstack-infra | 01:27 | |
| *** Ryan_Lane1 has quit IRC | 01:28 | |
| *** jyuso has joined #openstack-infra | 01:31 | |
| *** pcm_ has quit IRC | 01:33 | |
| *** salv-orlando has quit IRC | 01:34 | |
| *** Sukhdev has quit IRC | 01:35 | |
| *** yaguang has joined #openstack-infra | 01:38 | |
| *** nosnos has joined #openstack-infra | 01:38 | |
| *** gangil has joined #openstack-infra | 01:41 | |
| *** weshay has joined #openstack-infra | 01:44 | |
| *** baohua has joined #openstack-infra | 01:44 | |
| *** mbacchi has quit IRC | 01:48 | |
| *** xianghui has joined #openstack-infra | 01:49 | |
| *** ramashri has quit IRC | 01:50 | |
| *** bknudson has quit IRC | 01:59 | |
| *** camunoz has quit IRC | 02:02 | |
| *** armax has joined #openstack-infra | 02:04 | |
| *** sarob_ has joined #openstack-infra | 02:13 | |
| *** camunoz has joined #openstack-infra | 02:14 | |
| *** sarob_ has quit IRC | 02:18 | |
| *** armax has quit IRC | 02:20 | |
| *** otter768 has joined #openstack-infra | 02:22 | |
| *** arnaud has quit IRC | 02:22 | |
| *** yamahata has quit IRC | 02:37 | |
| openstackgerrit | Angus Lees proposed a change to openstack-infra/elastic-recheck: Add query for mysqldb + eventlet deadlock https://review.openstack.org/110531 | 02:37 |
| *** trinaths has joined #openstack-infra | 02:40 | |
| *** trinaths has quit IRC | 02:45 | |
| *** sarob_ has joined #openstack-infra | 02:46 | |
| *** otherwiseguy has joined #openstack-infra | 02:47 | |
| *** yamamoto_ has joined #openstack-infra | 02:48 | |
| *** sarob_ has quit IRC | 02:50 | |
| *** HenryG is now known as HenryG_afk | 02:53 | |
| *** Guest50025 has quit IRC | 02:55 | |
| *** SumitNaiksatam has joined #openstack-infra | 02:55 | |
| *** jgrimm has quit IRC | 02:55 | |
| *** harlowja is now known as harlowja_away | 02:57 | |
| *** medieval1 has joined #openstack-infra | 03:02 | |
| *** armax has joined #openstack-infra | 03:02 | |
| *** rainya has joined #openstack-infra | 03:03 | |
| *** jamielennox is now known as jamielennox|away | 03:05 | |
| *** medieval1_ has quit IRC | 03:06 | |
| *** yamamot__ has joined #openstack-infra | 03:17 | |
| *** ramashri has joined #openstack-infra | 03:18 | |
| *** morganfainberg is now known as morganfainberg_Z | 03:20 | |
| *** yamamoto_ has quit IRC | 03:20 | |
| *** ramashri has quit IRC | 03:22 | |
| *** morganfainberg_Z is now known as morganfainberg | 03:22 | |
| *** gyee has quit IRC | 03:23 | |
| *** arnaud has joined #openstack-infra | 03:23 | |
| *** arnaud__ has joined #openstack-infra | 03:24 | |
| *** rfolco has quit IRC | 03:25 | |
| *** ramashri has joined #openstack-infra | 03:25 | |
| *** weshay has quit IRC | 03:25 | |
| *** nosnos has quit IRC | 03:28 | |
| *** toabctl has joined #openstack-infra | 03:28 | |
| *** rainya has quit IRC | 03:31 | |
| *** rainya has joined #openstack-infra | 03:32 | |
| zaro | anteaya: FYI, we need to land one more change upstream https://gerrit-review.googlesource.com/#/c/58820/ | 03:32 |
| *** dimtruck is now known as zz_dimtruck | 03:32 | |
| *** rainya has quit IRC | 03:34 | |
| *** rainya has joined #openstack-infra | 03:34 | |
| *** terryw has joined #openstack-infra | 03:38 | |
| *** otherwiseguy has quit IRC | 03:40 | |
| *** otter768 has quit IRC | 03:46 | |
| *** alexandra has quit IRC | 03:47 | |
| *** armax has quit IRC | 03:48 | |
| *** terryw has quit IRC | 03:49 | |
| *** armax has joined #openstack-infra | 03:55 | |
| *** armax has quit IRC | 03:56 | |
| *** praneshp has quit IRC | 03:58 | |
| *** nosnos has joined #openstack-infra | 04:03 | |
| *** garyh has quit IRC | 04:10 | |
| *** amotoki has joined #openstack-infra | 04:14 | |
| *** Longgeek has joined #openstack-infra | 04:16 | |
| *** Longgeek has quit IRC | 04:18 | |
| *** Longgeek has joined #openstack-infra | 04:18 | |
| *** armax has joined #openstack-infra | 04:18 | |
| *** Longgeek has quit IRC | 04:19 | |
| *** Longgeek has joined #openstack-infra | 04:19 | |
| *** banix has quit IRC | 04:20 | |
| *** camunoz is now known as camunoz_away | 04:24 | |
| *** daya_k has joined #openstack-infra | 04:26 | |
| *** otherwiseguy has joined #openstack-infra | 04:27 | |
| *** trinaths has joined #openstack-infra | 04:29 | |
| *** daya_k_ has joined #openstack-infra | 04:30 | |
| *** mmaglana has joined #openstack-infra | 04:33 | |
| *** daya_k has quit IRC | 04:33 | |
| *** daya_k_ is now known as daya_k | 04:33 | |
| *** rainya has quit IRC | 04:35 | |
| *** terryw has joined #openstack-infra | 04:35 | |
| *** praneshp has joined #openstack-infra | 04:35 | |
| *** otherwiseguy has quit IRC | 04:37 | |
| *** camunoz_away is now known as camunoz | 04:39 | |
| *** garyh has joined #openstack-infra | 04:40 | |
| *** r1chardj0n3s is now known as r1chardj0n3s_afk | 04:42 | |
| *** nelsnelson has quit IRC | 04:42 | |
| *** terryw has quit IRC | 04:44 | |
| *** dims_ has joined #openstack-infra | 04:44 | |
| *** r1chardj0n3s_afk is now known as r1chardj0n3s | 04:47 | |
| *** LinStatSDR has joined #openstack-infra | 04:47 | |
| LinStatSDR | mornin | 04:50 |
| *** mmaglana has quit IRC | 04:52 | |
| *** ramashri has quit IRC | 04:52 | |
| *** mmaglana has joined #openstack-infra | 04:52 | |
| *** LinStatSDR has left #openstack-infra | 04:57 | |
| *** morganfainberg is now known as morganfainberg_Z | 04:59 | |
| zaro | jeblair: thanks for heads up on 91193, i'm on vacation until the 7th. Will babysit the merge when i get back. | 05:05 |
| zaro | jeblair: would appreciate you reasling python-jenkins ver 0.3.3 to pypi when you get a chance. it's the last release withou pbr. http://git.openstack.org/cgit/stackforge/python-jenkins/tag/?id=0.3.3 | 05:07 |
| zaro | *withou/without | 05:08 |
| *** ildikov has quit IRC | 05:09 | |
| zaro | jeblair: the tag build didn't make it onto tarballs because the build job failed to build wheels (due to it being a non-pbr project). http://logs.openstack.org/90/90b7ef732dcc813bf0a83301a8b21d65767dda64/post/python-jenkins-branch-tarball/1f6d7ec/console.html.gz | 05:13 |
| *** Ryan_Lane1 has joined #openstack-infra | 05:13 | |
| *** otherwiseguy has joined #openstack-infra | 05:14 | |
| *** dkehn_ has joined #openstack-infra | 05:22 | |
| *** dkehn has quit IRC | 05:22 | |
| *** dkehnx has quit IRC | 05:23 | |
| *** dkehn has joined #openstack-infra | 05:24 | |
| *** Ryan_Lane1 has quit IRC | 05:26 | |
| *** jcoufal has joined #openstack-infra | 05:28 | |
| *** gangil has quit IRC | 05:28 | |
| *** harlowja_at_home has joined #openstack-infra | 05:29 | |
| *** rlandy has joined #openstack-infra | 05:29 | |
| *** mriedem has joined #openstack-infra | 05:31 | |
| *** IvanBerezovskiy has joined #openstack-infra | 05:34 | |
| *** sarob_ has joined #openstack-infra | 05:36 | |
| *** Ryan_Lane has quit IRC | 05:36 | |
| *** samy has joined #openstack-infra | 05:38 | |
| *** sarob__ has joined #openstack-infra | 05:38 | |
| *** samy is now known as Guest99621 | 05:38 | |
| *** sarob_ has quit IRC | 05:41 | |
| *** sarob__ has quit IRC | 05:42 | |
| *** chandankumar has joined #openstack-infra | 05:44 | |
| *** pcrews has quit IRC | 05:45 | |
| *** Guest99621 has quit IRC | 05:47 | |
| *** pcrews has joined #openstack-infra | 05:48 | |
| *** bdpayne has joined #openstack-infra | 05:48 | |
| *** e0ne has joined #openstack-infra | 05:49 | |
| openstackgerrit | Christian Berendt proposed a change to openstack-infra/config: Add pep8 check/gate for stackforge/doc8 https://review.openstack.org/103863 | 05:51 |
| *** e0ne has quit IRC | 05:51 | |
| *** lukego has joined #openstack-infra | 05:55 | |
| *** yfried_ has joined #openstack-infra | 05:56 | |
| *** reed has quit IRC | 05:57 | |
| *** jgriffit1 has joined #openstack-infra | 06:01 | |
| *** terryw has joined #openstack-infra | 06:02 | |
| *** otherwiseguy has quit IRC | 06:02 | |
| *** dims_ has quit IRC | 06:03 | |
| *** harlowja_at_home has quit IRC | 06:04 | |
| *** jgriffith has quit IRC | 06:05 | |
| *** mriedem has quit IRC | 06:07 | |
| *** ildikov has joined #openstack-infra | 06:09 | |
| *** flaper87|afk is now known as flaper87 | 06:13 | |
| *** daya_k has quit IRC | 06:14 | |
| *** fifieldt_ is now known as fifieldt | 06:15 | |
| *** stevemar has quit IRC | 06:15 | |
| *** fifieldt has quit IRC | 06:19 | |
| *** YorikSar has quit IRC | 06:20 | |
| *** pcrews has quit IRC | 06:22 | |
| *** YorikSar has joined #openstack-infra | 06:22 | |
| *** jgriffit1 is now known as jgriffith | 06:24 | |
| *** terryw has quit IRC | 06:26 | |
| *** mpaolino has joined #openstack-infra | 06:29 | |
| *** e0ne has joined #openstack-infra | 06:29 | |
| *** e0ne has quit IRC | 06:29 | |
| *** bdpayne has quit IRC | 06:33 | |
| *** sarob_ has joined #openstack-infra | 06:36 | |
| *** arnaud has quit IRC | 06:37 | |
| *** arnaud__ has quit IRC | 06:37 | |
| *** sarob_ has quit IRC | 06:41 | |
| *** arnaud has joined #openstack-infra | 06:41 | |
| *** arnaud__ has joined #openstack-infra | 06:41 | |
| *** arnaud has quit IRC | 06:42 | |
| *** arnaud__ has quit IRC | 06:42 | |
| *** mmaglana has quit IRC | 06:51 | |
| *** mmaglana has joined #openstack-infra | 06:52 | |
| *** _nadya_ has joined #openstack-infra | 06:57 | |
| *** mmaglana has quit IRC | 06:57 | |
| *** armax has quit IRC | 07:04 | |
| *** ArxCruz has quit IRC | 07:11 | |
| *** sarob_ has joined #openstack-infra | 07:12 | |
| *** mpaolino_ has joined #openstack-infra | 07:13 | |
| *** mpaolino_ has quit IRC | 07:13 | |
| *** zns has joined #openstack-infra | 07:15 | |
| *** ArxCruz has joined #openstack-infra | 07:16 | |
| *** sarob_ has quit IRC | 07:17 | |
| *** amalagon has quit IRC | 07:19 | |
| *** KanagarajM has joined #openstack-infra | 07:19 | |
| *** ArxCruz has quit IRC | 07:21 | |
| *** _nadya_ has quit IRC | 07:22 | |
| *** fifieldt has joined #openstack-infra | 07:23 | |
| *** andreykurilin has joined #openstack-infra | 07:31 | |
| *** mmaglana has joined #openstack-infra | 07:34 | |
| *** luqas has joined #openstack-infra | 07:35 | |
| *** jistr has joined #openstack-infra | 07:35 | |
| *** sarob_ has joined #openstack-infra | 07:36 | |
| openstackgerrit | Dina Belova proposed a change to openstack-infra/config: Add job to run Tempest dsvm on MongoDB backend for Ceilometer https://review.openstack.org/110245 | 07:38 |
| *** mmaglana has quit IRC | 07:39 | |
| *** sarob_ has quit IRC | 07:41 | |
| *** tnovacik has joined #openstack-infra | 07:45 | |
| *** lxsli has joined #openstack-infra | 07:46 | |
| lxsli | So who do I need to kiss to get +A on https://review.openstack.org/#/c/108737/ ? :) | 07:46 |
| DinaBelova | folks, hello! I'm trying to add job to run Ceilometer tempest tests using mongodb backend... actually I'm going crazy as I really can't understand what I did wrong with the layout... gate-config-layout job is failing with "backend parameter missing to format gate-tempest-dsvm-ceilo-{backend}{branch-designator}" error - although it seems to me (he-he) that I did it ok.... | 07:50 |
| openstackgerrit | Dina Belova proposed a change to openstack-infra/config: Add job to run Tempest dsvm on MongoDB backend for Ceilometer https://review.openstack.org/110245 | 07:52 |
| *** e0ne has joined #openstack-infra | 07:54 | |
| *** jpich has joined #openstack-infra | 07:55 | |
| *** yamamot__ has quit IRC | 07:56 | |
| *** jgallard has joined #openstack-infra | 07:58 | |
| openstackgerrit | Dina Belova proposed a change to openstack-infra/config: Add job to run Tempest dsvm on MongoDB backend for Ceilometer https://review.openstack.org/110245 | 07:58 |
| *** ihrachyshka has joined #openstack-infra | 07:59 | |
| *** amuller has joined #openstack-infra | 07:59 | |
| openstackgerrit | A change was merged to openstack-infra/jeepyb: Added "no-gerrit" option. https://review.openstack.org/109069 | 07:59 |
| *** jedimike has joined #openstack-infra | 08:00 | |
| openstackgerrit | lokesh s proposed a change to openstack-infra/config: Adding new project https://review.openstack.org/110562 | 08:00 |
| *** jlibosva has joined #openstack-infra | 08:00 | |
| *** leopoldj has joined #openstack-infra | 08:01 | |
| *** skolekonov has joined #openstack-infra | 08:02 | |
| *** bauzas has joined #openstack-infra | 08:03 | |
| DinaBelova | question solved :) sorry for disturbing :) | 08:03 |
| *** bookwar has joined #openstack-infra | 08:08 | |
| *** derekh_ has joined #openstack-infra | 08:09 | |
| ttx | jeblair: on it | 08:12 |
| openstackgerrit | Dina Belova proposed a change to openstack-infra/config: Add job to run Tempest dsvm on MongoDB backend for Ceilometer https://review.openstack.org/110245 | 08:13 |
| *** roeyc has joined #openstack-infra | 08:16 | |
| *** yamamoto_ has joined #openstack-infra | 08:18 | |
| *** e0ne has quit IRC | 08:18 | |
| *** hashar has joined #openstack-infra | 08:18 | |
| openstackgerrit | A change was merged to openstack-infra/config: Publish specs jobs for *-specs https://review.openstack.org/106100 | 08:23 |
| openstackgerrit | A change was merged to openstack-infra/os-loganalyze: Bump hacking to 0.9.x series https://review.openstack.org/108307 | 08:27 |
| *** dmellado has joined #openstack-infra | 08:28 | |
| *** hashar has quit IRC | 08:29 | |
| openstackgerrit | A change was merged to openstack-infra/config: Collect tox logs from job runs https://review.openstack.org/106435 | 08:31 |
| *** dguerri`afk is now known as dguerri | 08:36 | |
| *** dguerri is now known as dguerri`afk | 08:36 | |
| *** dguerri`afk is now known as dguerri | 08:36 | |
| StevenK | SergeyLukjanov: Hai, can you poke at https://review.openstack.org/#/c/108817/1 , please? | 08:36 |
| *** sarob_ has joined #openstack-infra | 08:36 | |
| SergeyLukjanov | StevenK, hey, approved | 08:38 |
| StevenK | SergeyLukjanov: \o/ Thanks! | 08:38 |
| StevenK | greghaynes: ^ | 08:38 |
| openstackgerrit | Lucas Eznarriaga proposed a change to openstack-infra/config: Added advise to use dns over hardcoded ip in third_party https://review.openstack.org/104565 | 08:40 |
| *** sarob_ has quit IRC | 08:41 | |
| *** andreykurilin has quit IRC | 08:44 | |
| *** YorikSar has quit IRC | 08:44 | |
| *** YorikSar has joined #openstack-infra | 08:46 | |
| *** luqas has quit IRC | 08:47 | |
| *** hashar has joined #openstack-infra | 08:47 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add job to run Tempest dsvm on MongoDB backend for Ceilometer https://review.openstack.org/110245 | 08:49 |
| *** nosnos has quit IRC | 08:49 | |
| *** ZZelle has quit IRC | 08:51 | |
| *** ZZelle has joined #openstack-infra | 08:52 | |
| *** e0ne has joined #openstack-infra | 08:52 | |
| openstackgerrit | A change was merged to openstack-infra/config: New python-opentsdbclient repo https://review.openstack.org/108630 | 08:53 |
| *** pblaho has joined #openstack-infra | 08:56 | |
| *** daya_k has joined #openstack-infra | 08:58 | |
| *** hashar_ has joined #openstack-infra | 08:59 | |
| *** baohua has quit IRC | 09:00 | |
| *** baohua has joined #openstack-infra | 09:01 | |
| *** marzif has joined #openstack-infra | 09:01 | |
| openstackgerrit | A change was merged to openstack-infra/config: Force tripleo f20 novabm job to be novabm https://review.openstack.org/108817 | 09:02 |
| *** rcarrill` has joined #openstack-infra | 09:02 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/config: Fix specs upload job and use scp instead of ftp https://review.openstack.org/110573 | 09:02 |
| *** hashar has quit IRC | 09:02 | |
| *** afazekas has joined #openstack-infra | 09:03 | |
| openstackgerrit | A change was merged to openstack-infra/config: Gnocchi project changes monitoring added to the Ceilometer channel https://review.openstack.org/109601 | 09:03 |
| *** rcarrillocruz has quit IRC | 09:04 | |
| SergeyLukjanov | AJaeger, jhesketh, jeblair, fungi, clarkb, specs publishing job just failed with | 09:04 |
| SergeyLukjanov | Currently it fails with http://logs.openstack.org/7d/7dbd8ae848a3d5cf648ac06e4fd5a1a0cc8aeadf/post/sahara-specs-publish-specs/2e5c5c5/console.html#_2014-07-30_08_46_05_908 | 09:04 |
| SergeyLukjanov | 2014-07-30 08:46:05.908 | ERROR: Failed to find configuration named [specs.openstack.org] check the global configuration 2014-07-30 08:46:05.909 | Build step 'Send build artifacts over FTP' changed build result to UNSTABLE | 09:04 |
| SergeyLukjanov | so, I'm proposing to just use scp (as we always using it with statis.o.o) | 09:05 |
| SergeyLukjanov | https://review.openstack.org/110573 | 09:05 |
| SergeyLukjanov | jhesketh, if you're not sleeping, could you please take a look on it? | 09:05 |
| jhesketh | SergeyLukjanov: looking | 09:05 |
| jhesketh | SergeyLukjanov: see inline | 09:07 |
| jhesketh | SergeyLukjanov: Also, where are the credentials for scp stored? ie will this just work(tm) | 09:07 |
| *** KanagarajM2 has joined #openstack-infra | 09:07 | |
| *** hashar_ is now known as hashar | 09:07 | |
| SergeyLukjanov | jhesketh I think it's private key based | 09:08 |
| *** KanagarajM has quit IRC | 09:08 | |
| *** jp_at_hp has joined #openstack-infra | 09:08 | |
| SergeyLukjanov | jhesketh, but it'll work good anyway, we're doing exactly the same with logs | 09:08 |
| jhesketh | well it's not working at the moment, so it can't be worse | 09:08 |
| jhesketh | SergeyLukjanov: But looking at the error, it seems that the permissions needed to publish to specs.openstack.org aren't configured | 09:09 |
| jhesketh | so how do we know the scp ones are | 09:09 |
| SergeyLukjanov | jhesketh, specs.o.o is just a CNAME for static.o.o, so, IMO it's better to use the real place we're copying data to. | 09:10 |
| SergeyLukjanov | jhesketh, static.o.o is configured because we're copying logs with the same command - https://github.com/openstack-infra/config/blob/master/modules/openstack_project/files/jenkins_job_builder/config/macros.yaml#L431 | 09:11 |
| jhesketh | okay | 09:12 |
| jhesketh | SergeyLukjanov: okay, LGTm | 09:12 |
| SergeyLukjanov | jhesketh, thx | 09:13 |
| *** denis_makogon has joined #openstack-infra | 09:14 | |
| *** tnurlygayanov has quit IRC | 09:14 | |
| *** tnurlygayanov has joined #openstack-infra | 09:15 | |
| *** openstackgerrit has quit IRC | 09:16 | |
| *** openstackgerrit has joined #openstack-infra | 09:18 | |
| *** luqas has joined #openstack-infra | 09:19 | |
| *** rlandy has quit IRC | 09:21 | |
| *** baohua has quit IRC | 09:22 | |
| *** p5ntangle has joined #openstack-infra | 09:25 | |
| *** lttrl has joined #openstack-infra | 09:26 | |
| *** praneshp has quit IRC | 09:30 | |
| *** KanagarajM2 has quit IRC | 09:30 | |
| *** nosnos has joined #openstack-infra | 09:30 | |
| *** Longgeek_ has joined #openstack-infra | 09:31 | |
| openstackgerrit | Dina Belova proposed a change to openstack-infra/devstack-gate: Enable overriding of Ceilometer storage backend https://review.openstack.org/110247 | 09:32 |
| *** Hal has joined #openstack-infra | 09:32 | |
| *** KanagarajM has joined #openstack-infra | 09:32 | |
| openstackgerrit | lokesh s proposed a change to openstack-infra/config: Adding windows disk-image builder project https://review.openstack.org/110562 | 09:32 |
| *** Hal is now known as Guest70516 | 09:33 | |
| *** Longgeek has quit IRC | 09:34 | |
| ihrachyshka | hey all. I see some weird behavior when Change-Id is changed - no new review is created, instead patch is uploaded to existing review, and then it's not searchable thru the new Change-Id. F.e. see: https://review.openstack.org/#/c/83361/17 | 09:34 |
| ihrachyshka | I think it behaved differently before, right? | 09:35 |
| *** sarob_ has joined #openstack-infra | 09:36 | |
| *** jyuso has quit IRC | 09:36 | |
| openstackgerrit | Julien Marin-Frisonroche proposed a change to openstack-infra/git-review: Improve windows support for git-review --setup https://review.openstack.org/110331 | 09:37 |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish additional specs repositories https://review.openstack.org/109675 | 09:37 |
| *** sarob_ has quit IRC | 09:41 | |
| *** Steap__ has joined #openstack-infra | 09:43 | |
| *** SpamapS_ has joined #openstack-infra | 09:43 | |
| *** mwhudson_ has joined #openstack-infra | 09:43 | |
| *** slagle_ has joined #openstack-infra | 09:44 | |
| *** mikal_ has joined #openstack-infra | 09:44 | |
| *** blkperl_ has joined #openstack-infra | 09:44 | |
| *** KanagarajM has quit IRC | 09:44 | |
| *** derekh_ has quit IRC | 09:44 | |
| *** jpich has quit IRC | 09:44 | |
| *** ildikov has quit IRC | 09:44 | |
| *** lukego has quit IRC | 09:44 | |
| *** rkukura has quit IRC | 09:44 | |
| *** flaper87 has quit IRC | 09:44 | |
| *** mikal has quit IRC | 09:44 | |
| *** lxsli has quit IRC | 09:44 | |
| *** dansmith has quit IRC | 09:44 | |
| *** Steap has quit IRC | 09:44 | |
| *** sbadia has quit IRC | 09:44 | |
| *** mwhudson has quit IRC | 09:44 | |
| *** jd__ has quit IRC | 09:44 | |
| *** sdague has quit IRC | 09:44 | |
| *** vhoward has quit IRC | 09:44 | |
| *** SpamapS has quit IRC | 09:44 | |
| *** sc68cal has quit IRC | 09:44 | |
| *** clarkb has quit IRC | 09:44 | |
| *** devananda has quit IRC | 09:44 | |
| *** slagle has quit IRC | 09:44 | |
| *** jcooley has quit IRC | 09:44 | |
| *** blkperl has quit IRC | 09:44 | |
| *** briancline has quit IRC | 09:44 | |
| *** devanand1 has joined #openstack-infra | 09:44 | |
| *** jcooley has joined #openstack-infra | 09:44 | |
| *** lukego has joined #openstack-infra | 09:44 | |
| *** jd__ has joined #openstack-infra | 09:44 | |
| *** vhoward has joined #openstack-infra | 09:44 | |
| *** ildikov has joined #openstack-infra | 09:44 | |
| *** clarkb has joined #openstack-infra | 09:44 | |
| *** derekh_ has joined #openstack-infra | 09:44 | |
| *** dansmith has joined #openstack-infra | 09:44 | |
| *** jpich has joined #openstack-infra | 09:44 | |
| *** briancline has joined #openstack-infra | 09:44 | |
| *** lxsli has joined #openstack-infra | 09:44 | |
| *** sc68cal has joined #openstack-infra | 09:45 | |
| *** KanagarajM has joined #openstack-infra | 09:45 | |
| *** flaper87 has joined #openstack-infra | 09:45 | |
| *** sdague has joined #openstack-infra | 09:45 | |
| *** dansmith is now known as Guest17150 | 09:45 | |
| *** sbadia has joined #openstack-infra | 09:45 | |
| *** rkukura has joined #openstack-infra | 09:45 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish additional specs repositories https://review.openstack.org/109675 | 09:46 |
| *** yolanda has quit IRC | 09:46 | |
| *** unicell1 has joined #openstack-infra | 09:49 | |
| *** Steap__ is now known as Steap | 09:50 | |
| *** unicell has quit IRC | 09:52 | |
| *** BobBall has quit IRC | 09:55 | |
| openstackgerrit | Nicholas Randon proposed a change to openstack-infra/config: Add firewall traffic on port 27410 to 27510. https://review.openstack.org/108737 | 09:57 |
| *** e0ne has quit IRC | 10:02 | |
| *** e0ne has joined #openstack-infra | 10:02 | |
| *** BobBall has joined #openstack-infra | 10:08 | |
| openstackgerrit | Paul Bourke proposed a change to openstack-infra/reviewstats: Add --server param to reviewers https://review.openstack.org/110587 | 10:15 |
| openstackgerrit | Paul Bourke proposed a change to openstack-infra/reviewstats: Add missing spaces to --output usage message https://review.openstack.org/110597 | 10:20 |
| *** nosnos has quit IRC | 10:22 | |
| openstackgerrit | Paul Bourke proposed a change to openstack-infra/reviewstats: Add --server param to reviewers https://review.openstack.org/110587 | 10:22 |
| *** _nadya_ has joined #openstack-infra | 10:23 | |
| *** zns has quit IRC | 10:23 | |
| *** _nadya_ has quit IRC | 10:27 | |
| *** amotoki has quit IRC | 10:29 | |
| *** zns has joined #openstack-infra | 10:31 | |
| lxsli | Please could we get +2/+A on https://review.openstack.org/#/c/108737/ ? | 10:33 |
| *** jgallard has quit IRC | 10:34 | |
| *** sarob_ has joined #openstack-infra | 10:36 | |
| *** tnovacik has quit IRC | 10:40 | |
| *** sarob_ has quit IRC | 10:41 | |
| *** oanufriev has quit IRC | 10:42 | |
| *** zz_gondoi has quit IRC | 10:44 | |
| *** unicell1 has quit IRC | 10:45 | |
| *** unicell has joined #openstack-infra | 10:45 | |
| *** che-arne|2 has joined #openstack-infra | 10:48 | |
| *** che-arne has joined #openstack-infra | 10:48 | |
| *** che-arne has quit IRC | 10:48 | |
| *** leifmadsen has quit IRC | 10:49 | |
| *** Azher has joined #openstack-infra | 10:50 | |
| *** che-arne|2 is now known as che-arne | 10:54 | |
| *** leifmadsen has joined #openstack-infra | 10:54 | |
| *** gondoi has joined #openstack-infra | 10:54 | |
| *** ociuhandu has joined #openstack-infra | 10:55 | |
| *** Azher has quit IRC | 10:56 | |
| *** KanagarajM has quit IRC | 10:59 | |
| *** maxbit has quit IRC | 11:05 | |
| *** slagle_ is now known as slagle | 11:08 | |
| *** KanagarajM has joined #openstack-infra | 11:08 | |
| *** daya_k has quit IRC | 11:09 | |
| jhesketh | SergeyLukjanov: Are you going to be around to help if I merge 110247 | 11:11 |
| SergeyLukjanov | jhesketh, yup, I'll be here for at least 3 hours ;) | 11:12 |
| *** yamahata_ has quit IRC | 11:13 | |
| *** pcm_ has joined #openstack-infra | 11:21 | |
| *** yaguang has quit IRC | 11:23 | |
| *** dguerri is now known as dguerri`afk | 11:23 | |
| *** dguerri`afk is now known as dguerri | 11:23 | |
| *** zns has quit IRC | 11:23 | |
| *** zns has joined #openstack-infra | 11:24 | |
| fungi | SergeyLukjanov: jhesketh: the publisher destinations are configured in the jenkins masters themselves. the host you feed into the publisher is just a handle to the jenkins configuration where credentials, base directory and other parameters are manually specified | 11:28 |
| *** luqas has quit IRC | 11:29 | |
| jhesketh | SergeyLukjanov: Cool, will merge now | 11:29 |
| fungi | we use names for them which look like dns names, but from the perspective of jenkins they're arbitrary | 11:29 |
| jhesketh | fungi: okay, thanks | 11:30 |
| jhesketh | DinaBelova: Merging your ceilometer change if you're also around to help watch | 11:30 |
| SergeyLukjanov | jhesketh, ack | 11:30 |
| SergeyLukjanov | fungi, thx for clarification | 11:30 |
| SergeyLukjanov | fungi, could you please take a look at https://review.openstack.org/110573 | 11:31 |
| DinaBelova | jhesketh, /me here | 11:31 |
| fungi | you can find them configured at (for example) https://jenkins01.openstack.org/configure in the "scp repository hosts" secton | 11:31 |
| *** mmaglana has joined #openstack-infra | 11:32 | |
| fungi | so yes, the jobs were failing because there was no specs.openstack.org host defined in (in that case) the "publish over ftp" section on the masters where they ran | 11:34 |
| *** weshay has joined #openstack-infra | 11:35 | |
| *** mmaglana_ has joined #openstack-infra | 11:35 | |
| *** sarob_ has joined #openstack-infra | 11:36 | |
| *** mmaglana has quit IRC | 11:37 | |
| *** sarob_ has quit IRC | 11:41 | |
| *** KanagarajM has quit IRC | 11:43 | |
| fungi | ihrachyshka: i suspect this is due to newer gerrit's in-interface commit message editor. it looks like the commit message was edited through the webui and that didn't cause gerrit to see it as a separate change (note the change-id is still I49fd9afb40d4e5a5dade8e24da6f98170af27be2 in the detail table to the left of the commit message) | 11:48 |
| *** hashar has quit IRC | 11:49 | |
| ihrachyshka | fungi: probably you're right. so what should we do with those occurences? asking authors to reupload with git-gerrit? | 11:49 |
| *** mbacchi has joined #openstack-infra | 11:49 | |
| gilliard | https://review.openstack.org/#/c/108737/ is blocking some changes from being merged into tripleo-incubator, could someone take a look please? | 11:50 |
| fungi | ihrachyshka: probably, yes | 11:51 |
| *** rlandy has joined #openstack-infra | 11:51 | |
| *** dguerri is now known as dguerri`afk | 11:52 | |
| *** dguerri`afk is now known as dguerri | 11:52 | |
| *** adalbas has joined #openstack-infra | 11:54 | |
| *** dguerri is now known as dguerri`afk | 11:55 | |
| fungi | gilliard: that change is adjusting the firewall configuration of all our single-use slaves. at this point would it make sense to add a subclassed manifest and tie it just to the specific slave label which needs to run the job in question? (it would also help if the commit message identified the job or jobs this change is supporting, since i have no way of identifying which particular slaves are | 11:55 |
| fungi | impacted by the issue) | 11:55 |
| lxsli | fungi: that sounds good, no idea how to do that though! The fix is for the tripleo-ci pipeline | 11:58 |
| lxsli | hmmm maybe I can figure it out | 11:59 |
| *** xianghui has quit IRC | 12:01 | |
| fungi | lxsli: add a puppet manifest for something like an openstack_project::tripleo_slave which instantiates the single_use_slave class but overrides it with the firewall rules you need, and then adjust the tripleo-specific nodepool prep script to call that class instead of single_use_slave | 12:01 |
| lxsli | fungi: thanks, I'll have a go | 12:02 |
| gilliard | thanks for looking, guys | 12:02 |
| gilliard | fungi: Are you Jeremey Stanley on gerrit? | 12:03 |
| fungi | gilliard: yep | 12:03 |
| gilliard | cool - thanks again | 12:03 |
| fungi | gilliard: lxsli: i think separating these is a bit important in this case because as you'll note the existing firewall rules listed there are for internal sources within a devstack deployment (172.24.4.0/24) whereas i believe your 192.168.1.0/24 is an actual physical network within the tripleo provider clouds external to the node itself, correct? | 12:05 |
| fungi | in which case that network may be irrelevant or possibly have entirely different connotations within another non-tripleo cloud provider | 12:06 |
| lxsli | yep makes sense | 12:07 |
| derekh_ | lxsli: if that doesn't work, since this is tripleo specific, how about just adding iptables -A in toci_devtest.sh ? | 12:08 |
| derekh_ | or iptables -I | 12:09 |
| lxsli | I'll give the split a go | 12:11 |
| lxsli | but I'll certainly bear that in mind | 12:12 |
| *** rfolco has joined #openstack-infra | 12:14 | |
| *** leifmadsen has quit IRC | 12:16 | |
| *** leifmadsen has joined #openstack-infra | 12:16 | |
| *** dkranz has joined #openstack-infra | 12:22 | |
| *** HenryG_afk is now known as HenryG | 12:23 | |
| *** dguerri`afk is now known as dguerri | 12:24 | |
| DinaBelova | jhesketh, SergeyLukjanov - everything seems ok for now - http://status.openstack.org/zuul/ for 110247,2 change | 12:27 |
| DinaBelova | jhesketh, SergeyLukjanov, nothing suspicious yet :) | 12:27 |
| openstackgerrit | Nicholas Randon proposed a change to openstack-infra/config: Add firewall traffic on port 27410 to 27510. https://review.openstack.org/108737 | 12:29 |
| openstackgerrit | Nicholas Randon proposed a change to openstack-infra/config: Add firewall traffic on port 27410 to 27510. https://review.openstack.org/108737 | 12:30 |
| *** trinaths has quit IRC | 12:33 | |
| lxsli | fungi: I can't find a tripleo specific nodepool script? | 12:34 |
| *** sarob_ has joined #openstack-infra | 12:36 | |
| *** mfink has quit IRC | 12:37 | |
| openstackgerrit | litong01 proposed a change to openstack-infra/config: add channel #openstack-monasca to eavesdrop log server https://review.openstack.org/109109 | 12:38 |
| *** ArxCruz has joined #openstack-infra | 12:38 | |
| *** ildikov has quit IRC | 12:39 | |
| DinaBelova | jhesketh, hehe, failed for some strange reason | 12:39 |
| *** ildikov has joined #openstack-infra | 12:39 | |
| fungi | lxsli: at the moment http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/nodepool/scripts/prepare_node_tripleo.sh is calling prepare_node.sh which does the puppet apply call of the openstack_project::single_use_slave class, so i can imagine several ways you might rearrange that (perhaps add a MANIFEST=${8:-openstack_project::single_use_slave} and parameterize the | 12:40 |
| fungi | apply calls there, then pass it in your alternative class from the calling script | 12:40 |
| DinaBelova | jhesketh, all tempest tests are green and devstack was installed | 12:40 |
| *** pdmars has joined #openstack-infra | 12:40 | |
| DinaBelova | jhesketh, in console logs there are some strange traces - http://logs.openstack.org/47/110247/2/gate/gate-tempest-dsvm-full/c0c97eb/console.html | 12:40 |
| *** sarob_ has quit IRC | 12:41 | |
| *** yolanda has joined #openstack-infra | 12:44 | |
| *** aysyd has joined #openstack-infra | 12:49 | |
| gilliard | Here's a patch https://review.openstack.org/#/c/107057/6 - is says "1 comment" but when I click through there are no comments shown. | 12:50 |
| *** hashar has joined #openstack-infra | 12:50 | |
| *** timrc is now known as timrc-afk | 12:50 | |
| gilliard | The email I got from gerrit says the comment is attached to line #436 "please get rid of blank line", but ironically there isn't a line #436. | 12:50 |
| gilliard | Is this a bug in gerrit? | 12:51 |
| *** timrc-afk is now known as timrc | 12:52 | |
| *** cdub has joined #openstack-infra | 12:53 | |
| fungi | gilliard: probably. you might ask the commenter how they managed to add a comment for a nonexistent line (looks like they were talking about 435) | 12:54 |
| *** eharney has quit IRC | 12:54 | |
| fungi | i'm guessing they went through the api | 12:54 |
| *** nelsnelson has joined #openstack-infra | 12:54 | |
| *** yfried_ has quit IRC | 12:58 | |
| *** Guest70516 has quit IRC | 12:58 | |
| *** mfink has joined #openstack-infra | 12:59 | |
| *** pcm_ has quit IRC | 12:59 | |
| *** luqas has joined #openstack-infra | 13:00 | |
| *** Guest70516 has joined #openstack-infra | 13:01 | |
| AJaeger | SergeyLukjanov, thanks for fixing the specs upload with https://review.openstack.org/#/c/110573/ . Since jhesketh and myseld were confused about the "static.o.o", should we add a comment there? | 13:01 |
| * AJaeger learned something new about the infrastructure ;) | 13:01 | |
| AJaeger | SergeyLukjanov, I can also add the comment myself if you have no time right now so that this goes in quickly | 13:03 |
| *** mfink_ has joined #openstack-infra | 13:08 | |
| SergeyLukjanov | AJaeger, I'm not sure that it's really important - it's written in style used in many other places in config repo | 13:09 |
| SergeyLukjanov | AJaeger, the real mapping is the target path | 13:09 |
| *** mfink_ has quit IRC | 13:09 | |
| SergeyLukjanov | AJaeger, and Apache2 config | 13:09 |
| *** mfink_ has joined #openstack-infra | 13:09 | |
| AJaeger | SergeyLukjanov, then let's merge ;) | 13:09 |
| * AJaeger rereads the files and sees it - will +1 now. | 13:10 | |
| *** IvanBerezovskiy has left #openstack-infra | 13:11 | |
| *** ildikov has quit IRC | 13:11 | |
| SergeyLukjanov | fungi, what do you think about merging specs publishing fix? ^^ | 13:11 |
| *** mfink has quit IRC | 13:11 | |
| *** Hal has joined #openstack-infra | 13:13 | |
| *** Hal is now known as Guest53966 | 13:13 | |
| *** alexpilotti has joined #openstack-infra | 13:13 | |
| *** Guest70516 has quit IRC | 13:13 | |
| *** dustins has joined #openstack-infra | 13:14 | |
| *** alexpilotti_ has joined #openstack-infra | 13:19 | |
| *** alexpilotti has quit IRC | 13:19 | |
| *** alexpilotti_ is now known as alexpilotti | 13:19 | |
| *** Guest53966 has quit IRC | 13:20 | |
| *** lukego has quit IRC | 13:20 | |
| *** samy has joined #openstack-infra | 13:20 | |
| *** samy is now known as Guest42008 | 13:21 | |
| *** Guest53966 has joined #openstack-infra | 13:21 | |
| *** yamahata has joined #openstack-infra | 13:22 | |
| jhesketh | DinaBelova: That's odd.. I'm heading off now so I've removed the approve flag since it's failing | 13:22 |
| *** craigbr has joined #openstack-infra | 13:22 | |
| *** mriedem has joined #openstack-infra | 13:23 | |
| DinaBelova | jhesketh, ehm, ok then, although that were some glance client symbols errors there - and for some reason they appeared only in full-tempest job only in yhe gate - it was ok while check and all other jobs were ok as well | 13:23 |
| mriedem | are we supposed to be using some specific pypi mirror to get the alpha oslo packages (config, messaging, rootwrap) on a local tox run? | 13:24 |
| *** Guest53966 has quit IRC | 13:24 | |
| *** Hal_ has joined #openstack-infra | 13:24 | |
| gilliard | mriedem: pypi.python.org has 1.3.0.0a1 | 13:25 |
| mriedem | gilliard: they have 1.4.0.0a3 for oslo.config, which is what global-reqs wants, | 13:26 |
| mriedem | but a local tox rebuild of nova doesn't find that | 13:26 |
| mriedem | someone else in the ML confirmed the same | 13:26 |
| *** ildikov has joined #openstack-infra | 13:26 | |
| mriedem | when i tried just installing the latest oslo.config from pip it won't install the alpha package | 13:27 |
| mriedem | so wondering if there is a flag we need to use in tox.ini or something to get the alpha packages? | 13:27 |
| mriedem | ah looks like you need the --pre flag on the pip call | 13:28 |
| mriedem | so...why isn't that in everyone's tox.ini? | 13:29 |
| gilliard | I could install "oslo.config==1.4.0.0a3" without that flag | 13:30 |
| *** banix has joined #openstack-infra | 13:31 | |
| *** shashankhegde has joined #openstack-infra | 13:31 | |
| gilliard | Just tested it. Using a local devpi pointing at pypi.python.org | 13:31 |
| mriedem | yeah but i don't want to do that | 13:31 |
| mriedem | and tox isn't doing that | 13:31 |
| AJaeger | which tox versions are you using gillard and mriedem? | 13:31 |
| mriedem | 1.6.1 | 13:31 |
| gilliard | 1.7.2 | 13:32 |
| fungi | mriedem: you shouldn't need the --pre flag. tox turns that on by default (for questionable reasons). what pip version is ending up in your tox virtualenvs? if it's 1.4.1 you're running too old of a virtualenv (need 1.11 or later) | 13:32 |
| *** bknudson has joined #openstack-infra | 13:32 | |
| mriedem | fungi: pip 1.5.6 | 13:33 |
| *** mwagner_lap has joined #openstack-infra | 13:33 | |
| gilliard | same | 13:33 |
| fungi | mriedem: pip didn't start looking for wheels by default until 1.5, and keep in mind that virtualenv brings in its own embedded pip version rather than using the pip in your system context (under most circumstances) | 13:33 |
| fungi | mriedem: so that's the version of pip you have installed on your system, or the version of pip which virtualenv is installing in your tox venvs? | 13:34 |
| fungi | the latter is what matters | 13:34 |
| mriedem | fungi: i sourced .tox/py27/bin/activate and did pip show pip | 13:35 |
| mriedem | is there some other way? | 13:35 |
| fungi | mriedem: huh, yeah that's accurate then | 13:35 |
| *** dkranz has quit IRC | 13:35 | |
| *** afazekas has quit IRC | 13:35 | |
| fungi | mriedem: which project was this exposed under? | 13:35 |
| mriedem | fungi: nova | 13:35 |
| fungi | oh, you said nova | 13:35 |
| mriedem | and i'm using ubuntu precise | 13:35 |
| fungi | just rechecked scrollback | 13:36 |
| mriedem | fungi: just sent something to the oslo/nova ML on this too | 13:36 |
| mriedem | to separate from clarkb's tox hashseed thread | 13:36 |
| *** bknudson has quit IRC | 13:36 | |
| fungi | so, yeah, maybe the fact that nova still uses site packages in its tox.ini could be interacting strangely with your system pip | 13:36 |
| mriedem | i'll probably start bitching about libvirt-python next... | 13:36 |
| *** sarob_ has joined #openstack-infra | 13:36 | |
| fungi | libvirt-python should hopefully allow us to turn off site packages in nova's tox.ini now | 13:37 |
| fungi | since i think that was the only reason we had it enabled there | 13:37 |
| mriedem | https://review.openstack.org/#/c/105501/ | 13:37 |
| mriedem | yeah, well if you're on precise i think you have issues | 13:37 |
| mriedem | precise has libvirt 0.9.8 | 13:37 |
| mriedem | libvirt-python 1.2.5 appears to require libvirt >= 0.9.11 | 13:38 |
| *** dkliban has joined #openstack-infra | 13:38 | |
| fungi | oh, good point. you need newer libvirt to support the libvirt-python split | 13:38 |
| mriedem | i'm basically just waiting for someone to tell me 'f you, upgrade to trusty' :) | 13:38 |
| fungi | mriedem: well, we already did for upstream testing on juno/master ;) | 13:38 |
| *** mmaglana_ has quit IRC | 13:38 | |
| *** dkranz has joined #openstack-infra | 13:38 | |
| mriedem | i know, that's why i'm just waiting to be told to upgrade | 13:40 |
| mriedem | but gdi, it's early and i'm cranky | 13:40 |
| fungi | mriedem: you and me both | 13:40 |
| *** sarob_ has quit IRC | 13:41 | |
| *** maxbit has joined #openstack-infra | 13:41 | |
| mriedem | fungi: alright, well i've posted on both issues to the ML and i'll bring it up at the nova meetup today, thanks | 13:42 |
| *** kgiusti has joined #openstack-infra | 13:42 | |
| fungi | mriedem: so anyway, the story goes... virtualenv 1.11 will apply pip 1.5. that version of pip or later will install wheels by default without needing the --use-wheel option (which 1.4 does need, and 1.3 doesn't even know what wheels are). the --pre option says "install prerelease versions even if i don't ask for them" which is not really what you want, and pip 1.4 or later understand how to tell | 13:43 |
| fungi | prereleases apart from proper releases | 13:43 |
| *** Hal_ has quit IRC | 13:43 | |
| *** zz_dimtruck is now known as dimtruck | 13:43 | |
| *** Hal_ has joined #openstack-infra | 13:44 | |
| fungi | so in theory, with virtualenv>=1.11 you should be properly getting wheel-only prereleases within the version spec listed in your requirements.txt as long as the version listed in requirements.txt is a pre-release version | 13:44 |
| *** armax has joined #openstack-infra | 13:44 | |
| *** stevemar has joined #openstack-infra | 13:45 | |
| fungi | mriedem: oh! i just thoughth of one other possible problem... do you by any chance override your pip configuration to point to one of our mirrors? | 13:45 |
| fungi | mriedem: if your pip configuration is forcing use of our old mirror, it's abandoned and doesn't contain more recent packages | 13:46 |
| fungi | we should probably just delete it entirely at this point to make the broken more obvious | 13:46 |
| mriedem | fungi: nope, checked that last night, i only have use-mirrors=True | 13:47 |
| *** eharney has joined #openstack-infra | 13:47 | |
| fungi | oh, poop | 13:47 |
| fungi | do you have the pip.log from your failing run? | 13:47 |
| *** acfleury has joined #openstack-infra | 13:48 | |
| fungi | if you can stuff it up on paste.o.o i'll see if i can spot anything out of sorts | 13:48 |
| *** acfleury has quit IRC | 13:48 | |
| mriedem | gotta hit the shower before breakfast, i'll try later | 13:49 |
| *** Hal_ has quit IRC | 13:49 | |
| *** afazekas has joined #openstack-infra | 13:49 | |
| mriedem | *don't have to, but probably should... | 13:49 |
| *** wenlock has joined #openstack-infra | 13:49 | |
| *** freyes_ has joined #openstack-infra | 13:51 | |
| *** freyes_ has quit IRC | 13:51 | |
| *** amotoki has joined #openstack-infra | 13:52 | |
| armax | hi there, anyone who could help with these two: https://review.openstack.org/#/c/106901/ https://review.openstack.org/#/c/106902/ | 13:52 |
| armax | ? | 13:52 |
| armax | we’re pretty close to finishing Neutron DVR off and we can have an experimental job we can test it with it would be super helpful | 13:53 |
| *** adalbas has quit IRC | 13:54 | |
| armax | I know you guys are busy :) | 13:54 |
| openstackgerrit | David Pineau proposed a change to openstack-infra/infra-manual: Write rough outline for the 3rd-party ci manual https://review.openstack.org/110651 | 13:54 |
| joa | anteaya: krtaylor: ^ | 13:54 |
| *** jistr has quit IRC | 13:54 | |
| *** dkehn_ is now known as dkehnx | 13:55 | |
| armax | thanks fungi! | 13:55 |
| *** jistr has joined #openstack-infra | 13:55 | |
| *** Hal_ has joined #openstack-infra | 13:57 | |
| *** dmsimard_away is now known as dmsimard | 13:57 | |
| fungi | armax: see comment on the config change | 13:57 |
| joa | anteaya: the first patch is a bit bigger than discussed, but I wanted to lay out the whole thing before we could discuss it and improve :) | 13:57 |
| armax | fungi: I am on it! | 13:57 |
| *** adalbas has joined #openstack-infra | 13:57 | |
| fungi | armax: it's a soft -1, mainly trying to figure out how ^(?!stable/(?:havana|icehouse)).*$ differs from ^(?!stable/(havana|icehouse)).*$ | 13:58 |
| armax | fungi: agreed, I followed Andreas’ suggestion on PS3 | 13:58 |
| fungi | specifically the (?:...) isn't a convention i've used | 13:59 |
| *** dimtruck is now known as zz_dimtruck | 13:59 | |
| armax | fungi: I am still wrapping my head around how infra works :) | 13:59 |
| fungi | so mainly trying to avoid digging back through the python regex chapter | 13:59 |
| armax | fungi: but you have a point, if it works for other jobs, we might as well adopt the same regex | 13:59 |
| *** dustins has quit IRC | 14:00 | |
| fungi | armax: the secret is that we don't work... we just program an army of bots to do all the work for us so we can drink beer and travel | 14:00 |
| armax | fungi: oh, love it! can I join? :) | 14:00 |
| fungi | yes, there's not even an application fee ;) | 14:00 |
| *** dustins has joined #openstack-infra | 14:01 | |
| armax | fungi: great, I’ll bring a couple of six-packs | 14:01 |
| lxsli | fungi: () is a capturing group, (?:) is non-capturing | 14:02 |
| fungi | lxsli: thanks! | 14:02 |
| armax | fungi: gate-designate-devstack-dsvm has the same regex expression | 14:02 |
| armax | fungi: so not 100% sure how to go about this | 14:03 |
| fungi | armax: i see, nifty. now the only remaining question i have is whether we're likely to confuse contributors with too much regex proliferation accomplishing (in most cases) the same outcome | 14:03 |
| lxsli | fungi: I'm going to need your help on my next patch so if I can repay with regex knowledge that's great :) | 14:03 |
| *** mriedem has quit IRC | 14:03 | |
| fungi | lxsli: my mind has been muddied by years of juggling posix regex, pcre and even (eww) cisco regular expressions | 14:04 |
| * fungi can't keep them all straight any longer | 14:04 | |
| armax | fungi: fair enough, it does look like gate-designate-devstack-dsvm and the newly introduced check-tempest-dsvm-neutron-dvr-enabled would be the only ones following this pattern | 14:05 |
| openstackgerrit | A change was merged to openstack/requirements: Add new library oslo.utils https://review.openstack.org/110380 | 14:05 |
| lxsli | I have enough trouble with POSIX + Ex | 14:05 |
| *** primeministerp has joined #openstack-infra | 14:06 | |
| fungi | armax: your goal was simply to run this on any branch which is neither stable/havana nor stable/icehouse correct? | 14:06 |
| armax | fungi: master and stable branches going forward, yes | 14:06 |
| fungi | AJaeger: if you're around, did you recommend a non-capturing group in the regex on 106902 for a particular reason? | 14:07 |
| AJaeger | fungi, good catch - I wonder as well now. | 14:08 |
| openstackgerrit | Alexis Lee proposed a change to openstack-infra/config: Add firewall traffic on port 27410 to 27510 https://review.openstack.org/110653 | 14:08 |
| fungi | i definitely don't want to push for consistency with existing expressions if they're subtly wrong and this one is better, but otherwise we should probably try to keep them consistent to avoid confusing people (for example me!) | 14:09 |
| *** dkranz has quit IRC | 14:09 | |
| AJaeger | fungi, seems I copied without realizing that this subtile difference. | 14:09 |
| AJaeger | I just referenced the first regex ... | 14:09 |
| armax | fungi, AJaeger I think this is mainly a matter of consistency I think both expressions correctly filter the branch | 14:09 |
| armax | correct? | 14:10 |
| fungi | AJaeger: okay, no problem. thanks! we likely want to fix the one existing occurrence of that in the config to also be consistent | 14:10 |
| jeblair | AJaeger, mordred: i have a solution to the problem we discussed in meeting: we put a marker file at the root of any rsync directory, and use that to build an exclusion list (on the publishing server) before running the rsync command | 14:10 |
| fungi | armax: yes, agreed | 14:10 |
| jeblair | AJaeger, mordred: i'll write it up in more detail in a spec | 14:10 |
| AJaeger | jeblair, I wsa thinking of a spec as well, want to start later on it. IF you give it a first go, I can enhance with the doc view. | 14:11 |
| fungi | jeblair: just passing it into --exclude-from? | 14:11 |
| jeblair | fungi: yep | 14:11 |
| fungi | that should be clean | 14:11 |
| AJaeger | Yeah, we can enhance the tools to output that.... | 14:11 |
| *** pcrews has joined #openstack-infra | 14:11 | |
| AJaeger | But let me write up the examples to see whether it handles everything | 14:11 |
| openstackgerrit | David Pineau proposed a change to openstack-infra/infra-manual: Write rough outline for the 3rd-party ci manual https://review.openstack.org/110651 | 14:11 |
| jeblair | AJaeger: sounds good; i'll do a first draft soon | 14:12 |
| armax | fungi, AJaeger: looking closely we have 5 instances of (havana|icehouse) and 2 instances of (?:havana|icehouse) | 14:12 |
| *** esker has joined #openstack-infra | 14:12 | |
| AJaeger | jeblair, thanks! | 14:12 |
| armax | fungi: at this point I’d flip back to the more common one | 14:12 |
| armax | and we can fix the consistency in a follow-up patch | 14:12 |
| * AJaeger is currently chairing a call - have to disappear again ;) | 14:12 | |
| jeblair | fungi, AJaeger: we actually need the swift-upload script to generate a manifest file so that we know what to download from swift anyway, could probably just leave that in place. | 14:13 |
| jeblair | AJaeger: np. i'll put it all in the spec :) | 14:13 |
| AJaeger | jeblair, thanks! | 14:13 |
| openstackgerrit | Armando Migliaccio proposed a change to openstack-infra/config: Add experimental job Neutron DVR testing on single node https://review.openstack.org/106902 | 14:13 |
| *** sabeen has quit IRC | 14:14 | |
| *** wenlock has quit IRC | 14:14 | |
| *** gondoi is now known as zz_gondoi | 14:14 | |
| fungi | jeblair: yep, should work great | 14:15 |
| *** zz_dimtruck is now known as dimtruck | 14:18 | |
| *** jistr has quit IRC | 14:18 | |
| *** jistr has joined #openstack-infra | 14:19 | |
| openstackgerrit | Armando Migliaccio proposed a change to openstack-infra/config: Use same regex expression to match stable branches https://review.openstack.org/110656 | 14:19 |
| *** jgrimm has joined #openstack-infra | 14:19 | |
| *** shashankhegde has quit IRC | 14:19 | |
| armax | fungi, AJaeger: thanks for your help, when you have some time have a look at https://review.openstack.org/110656 | 14:20 |
| *** afazekas has quit IRC | 14:20 | |
| *** rcarrillocruz has joined #openstack-infra | 14:21 | |
| *** dkranz has joined #openstack-infra | 14:21 | |
| fungi | armax: thanks a bunch. i've also added Kiall to the review just in case he did that for some subtle reason i'm not seeing | 14:22 |
| armax | fungi: thank you! | 14:22 |
| armax | sure | 14:22 |
| *** rcarrill` has quit IRC | 14:23 | |
| *** dimtruck is now known as zz_dimtruck | 14:23 | |
| *** zul has joined #openstack-infra | 14:23 | |
| *** signed8bit has joined #openstack-infra | 14:24 | |
| fungi | armax: ha! see jeblair's foreshadowing in his comment to https://review.openstack.org/102864 | 14:24 |
| *** cnesa7 has joined #openstack-infra | 14:25 | |
| fungi | and i apparently approved it | 14:25 |
| * jeblair polishes crystal ball | 14:25 | |
| phschwartz | fungi: I will check with them again this morning. Ops was suppose to be taking care of it. | 14:26 |
| *** otherwiseguy has joined #openstack-infra | 14:26 | |
| fungi | phschwartz: i'll check again as well. they may have at this point | 14:26 |
| openstackgerrit | Boris Pavlovic proposed a change to openstack-infra/config: Cleanup rally & osprofiler jobs https://review.openstack.org/110662 | 14:26 |
| armax | fungi: ah! | 14:26 |
| fungi | phschwartz: nope. count is now at exactly 100 nodes in error state | 14:26 |
| openstackgerrit | A change was merged to openstack-infra/config: Fix specs upload job and use scp instead of ftp https://review.openstack.org/110573 | 14:27 |
| armax | fungi: it sounds like we looked at it the other way around :) | 14:27 |
| *** Hal has joined #openstack-infra | 14:30 | |
| *** Hal_ has quit IRC | 14:30 | |
| *** Hal is now known as Guest53224 | 14:30 | |
| *** homeless has joined #openstack-infra | 14:31 | |
| *** afazekas has joined #openstack-infra | 14:33 | |
| openstackgerrit | Boris Pavlovic proposed a change to openstack-infra/config: Cleanup rally & osprofiler jobs https://review.openstack.org/110662 | 14:35 |
| *** sarob_ has joined #openstack-infra | 14:36 | |
| fungi | so it looks like once we're properly expiring docs-draft contents, we can probably shuffle it onto a much smaller lv: http://cacti.openstack.org/cacti/graph.php?action=view&local_graph_id=718&rra_id=all | 14:38 |
| *** zz_dimtruck is now known as dimtruck | 14:39 | |
| fungi | (note that it's still in the process of cleaning up, so could go considerably smaller than it's at now even) | 14:39 |
| *** wenlock has joined #openstack-infra | 14:39 | |
| fungi | alternatively, we could up the retention to more than 3 weeks | 14:39 |
| *** dangers_away is now known as dangers | 14:39 | |
| *** sarob_ has quit IRC | 14:41 | |
| lxsli | please could someone that knows Puppet give some feedback on https://review.openstack.org/110653 ? I know it's wrong but I don't want to learn all Puppet syntax just for this one change :/ | 14:43 |
| openstackgerrit | Thierry Carrez proposed a change to openstack-infra/release-tools: spec2bp v2 -- support in-review specs https://review.openstack.org/108041 | 14:43 |
| *** zul has quit IRC | 14:47 | |
| *** signed8bit has quit IRC | 14:50 | |
| *** chandankumar_ has joined #openstack-infra | 14:50 | |
| *** kgiusti has quit IRC | 14:51 | |
| *** thedodd has joined #openstack-infra | 14:52 | |
| *** mfer has joined #openstack-infra | 14:52 | |
| *** kgiusti has joined #openstack-infra | 14:52 | |
| *** chandankumar has quit IRC | 14:53 | |
| *** datsun180b has joined #openstack-infra | 14:53 | |
| *** jgallard has joined #openstack-infra | 14:53 | |
| *** signed8bit has joined #openstack-infra | 14:54 | |
| *** rainya has joined #openstack-infra | 14:54 | |
| *** juzuluag has joined #openstack-infra | 14:56 | |
| *** otherwiseguy has quit IRC | 14:57 | |
| *** rwsu has joined #openstack-infra | 14:58 | |
| *** craigbr has quit IRC | 15:00 | |
| *** bdpayne has joined #openstack-infra | 15:00 | |
| *** reed has joined #openstack-infra | 15:00 | |
| *** rcarrill` has joined #openstack-infra | 15:01 | |
| *** flaper87 is now known as flwang_ | 15:01 | |
| *** flwang_ is now known as flaper87 | 15:01 | |
| *** rcarrillocruz has quit IRC | 15:03 | |
| *** _nadya_ has joined #openstack-infra | 15:06 | |
| *** vhoward has left #openstack-infra | 15:09 | |
| *** sabeen1 has joined #openstack-infra | 15:09 | |
| *** zz_gondoi is now known as gondoi | 15:10 | |
| cnesa7 | Hi all, we have a blocking bug on the HyperV CI, devstack fails during neutron database creation. stack trace here -> http://paste.openstack.org/show/88896/. Is there a workaround? | 15:10 |
| *** roeyc has quit IRC | 15:10 | |
| *** skolekonov has quit IRC | 15:11 | |
| *** ramashri has joined #openstack-infra | 15:12 | |
| *** signed8bit has quit IRC | 15:12 | |
| *** MarkAtwood has joined #openstack-infra | 15:12 | |
| *** signed8bit has joined #openstack-infra | 15:13 | |
| *** david-lyle has joined #openstack-infra | 15:14 | |
| *** e0ne has quit IRC | 15:15 | |
| *** afazekas has quit IRC | 15:17 | |
| *** zul has joined #openstack-infra | 15:17 | |
| *** bdpayne has quit IRC | 15:19 | |
| *** sballe_ has joined #openstack-infra | 15:19 | |
| *** gokrokve has joined #openstack-infra | 15:20 | |
| fungi | cnesa7: alembic seems to think your migrations are branched... i'm not super familiar with it though. have you looked at the alembic documentation for pointers on how to debug that? | 15:21 |
| *** mfer has quit IRC | 15:21 | |
| *** sballe has joined #openstack-infra | 15:22 | |
| *** david-lyle has quit IRC | 15:22 | |
| *** david-lyle has joined #openstack-infra | 15:23 | |
| stevemar | AJaeger, darn, doesn't look like it worked, https://review.openstack.org/#/c/110658/ just merged, but http://specs.openstack.org/openstack/keystone-specs keeps coming up empty | 15:24 |
| stevemar | oh wait a minute... | 15:24 |
| *** vhoward has joined #openstack-infra | 15:24 | |
| stevemar | i think it just needed a few minutes to upload the data | 15:25 |
| AJaeger | http://specs.openstack.org/openstack/keystone-specs/doc/build/html/ works | 15:25 |
| AJaeger | A bit deeper than I expected ;( | 15:25 |
| stevemar | lol yes, i jumped the gun | 15:25 |
| stevemar | yeah | 15:25 |
| stevemar | fixable | 15:25 |
| *** ihrachyshka_ has joined #openstack-infra | 15:25 | |
| *** ihrachyshka has quit IRC | 15:26 | |
| *** Guest42008 has quit IRC | 15:26 | |
| AJaeger | fungi, SergeyLukjanov: https://review.openstack.org/#/c/110573 removed the remove-prefix: 'doc/build/html' line that we added there on purpose | 15:26 |
| fungi | AJaeger: out of curiosity, do you know the reasoning behind http://specs.openstack.org/robots.txt ? do we specifically not want search engines indexing our specs? | 15:26 |
| AJaeger | What's the proper way to do this with scp? | 15:26 |
| joa | Small question about openstack-infra/config | 15:26 |
| *** Guest42008 has joined #openstack-infra | 15:26 | |
| AJaeger | fungi, that's a question for stevemar ;) | 15:26 |
| joa | I'm looking to understand how to configure a dsvm-tempest-full for a cinder CI; annd I cannot find the template definition of this job ? | 15:27 |
| *** bknudson has joined #openstack-infra | 15:28 | |
| *** gyee has joined #openstack-infra | 15:28 | |
| *** sballe has quit IRC | 15:29 | |
| fungi | joa: http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/jenkins_job_builder/config/devstack-gate.yaml#n94 | 15:29 |
| stevemar | fungi, my mistake on that one, i was using 'docs' as a guide, and must not have noticed that it was docs-draft that had the robots entry | 15:30 |
| stevemar | fungi, want me to toss up a patch to remove that? | 15:30 |
| joa | fungi: okay, so sometimes the branch-designator is empty/undefined ? | 15:31 |
| fungi | stevemar: maybe wait until we get the publishing ironed out | 15:31 |
| fungi | joa: usually branch-designator is empty | 15:31 |
| joa | is it given by zuul ? | 15:32 |
| *** trinaths has joined #openstack-infra | 15:32 | |
| fungi | joa: it's mainly used to give branch-specific jobs unique names, though also sometimes (ab)used to make multiple versions of the same job so they can be added to different pipelines and filtered to different triggering branches | 15:32 |
| joa | I'm trying to understand the details of jenkins template jobs + zuul to be able to configure my ci :p | 15:32 |
| AJaeger | stevemar, I'm doing a patch for the location now... | 15:33 |
| fungi | joa: it's a parameter you'll pass in when instantiating the template by adding it or a job group containing it to a project | 15:33 |
| stevemar | AJaeger, awesome, i'll toss one up for robots then | 15:33 |
| fungi | joa: in the jjb configuration, not in zuul's layout | 15:33 |
| redrobot | hi guys, any core devs around to give https://review.openstack.org/#/c/109398/ a look? | 15:33 |
| *** Sukhdev has joined #openstack-infra | 15:34 | |
| joa | fungi: ok I think I get it :) | 15:34 |
| redrobot | it's a one liner, so it shouldn't take more than a few seconds. :D | 15:34 |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish *-specs better https://review.openstack.org/110682 | 15:35 |
| AJaeger | fungi, could you tell me whether the patch does what it should do? See the commit message. | 15:35 |
| *** [1]trinaths has joined #openstack-infra | 15:36 | |
| fungi | stevemar: AJaeger: looking over how the publisher for that is expanded to the configuration in jenkins itself, it looks like there's no way to re-root the source to exclude parts of the path if you're keeping the hierarchy, and of you don't keep the hierarchy then all files wind up in one flat target directory with no subdirectory structure | 15:36 |
| AJaeger | That's sad ;( | 15:36 |
| AJaeger | so, should we go back to ftp publishing - ftp allows this... | 15:37 |
| openstackgerrit | A change was merged to openstack-infra/release-tools: Bump hacking to 0.9.x series https://review.openstack.org/108666 | 15:38 |
| AJaeger | so, basically reverting https://review.openstack.org/#/c/110573/1/modules/openstack_project/files/jenkins_job_builder/config/specs-jobs.yaml - but using static insted of specs.o.o for sit? | 15:38 |
| AJaeger | fungi, would this work with the ftp publisher - or can't we use ftp publishing for static? | 15:39 |
| *** trinaths has quit IRC | 15:39 | |
| *** [1]trinaths is now known as trinaths | 15:39 | |
| fungi | AJaeger: static.o.o doesn't run an ftp service as far as i'm aware, though i can double-check | 15:39 |
| *** mmaglana has joined #openstack-infra | 15:39 | |
| AJaeger | and scp can change paths? That sounds like fun ;( | 15:40 |
| AJaeger | In that case we have to leave it as is and create some nice overview page ;) | 15:40 |
| fungi | AJaeger: one workaround would be to make the doc build move the doc/build/html to {github-org}/{name} and then use that as the source in the publisher with a target of specs/ | 15:41 |
| *** amuller has quit IRC | 15:41 | |
| AJaeger | fungi, yes, that would work | 15:41 |
| fungi | stevemar: AJaeger: also, we probably want to disable copy after failure (it's useful for logs, but not so much when trying to publish documentation, unless i'm missing something) | 15:42 |
| *** james_li has joined #openstack-infra | 15:42 | |
| *** otherwiseguy has joined #openstack-infra | 15:43 | |
| openstackgerrit | Steve Martinelli proposed a change to openstack-infra/config: Remove the robots entry from specs.openstack.org https://review.openstack.org/110689 | 15:43 |
| stevemar | fungi, AJaeger ^ | 15:43 |
| *** mmaglana has quit IRC | 15:44 | |
| stevemar | fungi, AJaeger ++ on moving the content from doc/build/html a few levels up - makes sense | 15:45 |
| * AJaeger is on it... | 15:45 | |
| *** chandankumar_ has quit IRC | 15:45 | |
| *** salv-orlando has joined #openstack-infra | 15:46 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish *-specs better https://review.openstack.org/110682 | 15:46 |
| AJaeger | fungi, can this work? | 15:46 |
| openstackgerrit | Ala Rezmerita proposed a change to openstack-infra/config: Add swiftpolicy middleware project on StackForge https://review.openstack.org/110694 | 15:47 |
| AJaeger | stevemar, fungi : I've updated https://review.openstack.org/110682 | 15:47 |
| *** jlibosva has quit IRC | 15:48 | |
| fungi | AJaeger: not quite, i'll suggest a tweak inline. you're ending up copying an extra directory level you don't want there | 15:49 |
| *** bookwar has quit IRC | 15:49 | |
| *** gyee has quit IRC | 15:51 | |
| *** gokrokve_ has joined #openstack-infra | 15:51 | |
| *** dims_ has joined #openstack-infra | 15:52 | |
| *** dims_ is now known as dims | 15:52 | |
| stevemar | AJaeger, any reason we do a copy and not a move? | 15:52 |
| fungi | stevemar: AJaeger: yes, mv also seems fine for this (and perhaps preferable to cp) | 15:53 |
| * AJaeger will update, thanks for reviewing! | 15:53 | |
| stevemar | np! | 15:53 |
| *** mfer has joined #openstack-infra | 15:54 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish *-specs better https://review.openstack.org/110682 | 15:54 |
| *** mmaglana has joined #openstack-infra | 15:54 | |
| *** gokrokve has quit IRC | 15:55 | |
| AJaeger | next iteration... | 15:55 |
| fungi | AJaeger: lgtm | 15:55 |
| *** gokrokve_ has quit IRC | 15:55 | |
| AJaeger | stevemar, fungi : I'll on my way home now and will followup later if you notice anything. Feel free to edit yourself if needed. | 15:55 |
| AJaeger | Great, fungi ! | 15:55 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/infra-specs: Remove test spec https://review.openstack.org/110701 | 15:55 |
| * AJaeger loves team scripting with fungi and stevemar ;) | 15:56 | |
| stevemar | AJaeger, cool thanks! | 15:56 |
| stevemar | haha | 15:56 |
| stevemar | yay team scripting | 15:56 |
| *** hemna_ is now known as hemna | 15:56 | |
| joa | Is there an easy way to configure a common project/trigger to run two (VEry) similar jkobs on two different slaves ? | 15:56 |
| *** nkubota_ has joined #openstack-infra | 15:56 | |
| AJaeger | I know pair programming but didn't know how to call what we did ;) | 15:56 |
| cnesa7 | fungi: thank you | 15:57 |
| joa | I saw the "node: 'slave1 || slave2'" syntax for a given job, but I do not know what it means exactly | 15:57 |
| fungi | joa: you specify the "node" parameter in each job to be the name or label of the slaves where you want each to run | 15:57 |
| joa | is it a "list of available node" or does the job have to run on both ? | 15:57 |
| fungi | joa: that's a list of nodes where the job *can* run but any one invocation will just run on the first available | 15:57 |
| openstackgerrit | Julien Marin-Frisonroche proposed a change to openstack-infra/git-review: Improve windows support for git-review --setup https://review.openstack.org/110331 | 15:58 |
| joa | ok; so I must define two jobs templates if I want to run the same job on two different nodes ? | 15:58 |
| *** bdpayne has joined #openstack-infra | 15:58 | |
| *** shashankhegde has joined #openstack-infra | 15:58 | |
| *** dustins has quit IRC | 15:58 | |
| joa | (btw, what I'm trying to setup is a cinder tempest-dsvm-full on two different slaves for two different drivers) | 15:58 |
| *** vermarah_ has joined #openstack-infra | 15:58 | |
| fungi | joa: it depends on whether the nodes are actually distinct in function, or if they're interchangeable for you | 15:58 |
| *** rcarrillocruz has joined #openstack-infra | 15:59 | |
| openstackgerrit | Ala Rezmerita proposed a change to openstack-infra/config: Add swiftpolicy middleware project on StackForge https://review.openstack.org/110694 | 15:59 |
| joa | they are distinct, since they each have their specific vendor environment | 15:59 |
| *** ildikov has quit IRC | 15:59 | |
| joa | ideally, I'd want a report/log for each of those nodes | 15:59 |
| fungi | joa: if the machines are configured differently outside the job itself, to support the two different jobs, then you want to just specify the name of the slave as the node for the job to which you're matching it | 15:59 |
| *** marcoemorais has joined #openstack-infra | 16:00 | |
| *** _nadya_ has quit IRC | 16:00 | |
| joa | okay; can I specify the same job twice, with a different node each time ? | 16:00 |
| *** marun has quit IRC | 16:00 | |
| *** p5ntangl_ has joined #openstack-infra | 16:00 | |
| fungi | joa: so your job template might have node: '{node}' and then in each separate instantiation of that template you'd set the node parameter to what it needs to be for that particular job | 16:00 |
| *** p5ntangl_ has quit IRC | 16:01 | |
| *** e0ne has joined #openstack-infra | 16:01 | |
| *** rcarrill` has quit IRC | 16:01 | |
| *** mdenny has joined #openstack-infra | 16:01 | |
| joa | we agree that the isntanciation is done within a jjb project ? | 16:01 |
| *** dustins has joined #openstack-infra | 16:01 | |
| fungi | joa: or it may be simpler to just use two separate non-templated job definitions with a different node in each | 16:01 |
| fungi | joa: yeah, if you're instantiating a template you'd usually do that in a jjb project configuration | 16:01 |
| joa | Actually, I'd like to keep my job template definitions the closest possible from openstack-infra/config :p | 16:02 |
| *** marun has joined #openstack-infra | 16:02 | |
| *** p5ntangle has quit IRC | 16:03 | |
| *** gangil has joined #openstack-infra | 16:04 | |
| joa | fungi: other question: the jjb lives on the jenkins master only or also on the slaves ? (I'm thinking of the gerrit-git-prep builder, which refers a script slave-side) | 16:05 |
| joa | (sorry about this question storm btw :D) | 16:05 |
| *** gyee has joined #openstack-infra | 16:05 | |
| *** sabeen2 has joined #openstack-infra | 16:05 | |
| asselin | joa, jjb is only on the master | 16:06 |
| joa | okay, thanks | 16:06 |
| joa | asselin: you might be interested in following what I started for the third_party.rst in infra/manuals | 16:07 |
| *** markmcclain has joined #openstack-infra | 16:07 | |
| joa | https://review.openstack.org/#/c/110651/ | 16:07 |
| joa | if only to add your own feedback | 16:07 |
| fungi | joa: jjb is a tool which runs on the jenkins master to translate a bunch of templated yaml files into xml job configs which get loaded in via the jenkins api over the loopback interface | 16:07 |
| joa | okay | 16:07 |
| *** sabeen1 has quit IRC | 16:07 | |
| joa | it was one of the few unclear bits for me in the whole architecture :p | 16:07 |
| joa | thanks for clarifying it | 16:08 |
| fungi | joa: the jenkins master may in turn apply some of the configuration which has been loaded into it as things which the slaves should run during the course of a job (particularly "shell" builders) | 16:08 |
| joa | ok | 16:09 |
| jeblair | joa: have you seen http://ci.openstack.org/running-your-own.html ? | 16:09 |
| *** gyee has quit IRC | 16:09 | |
| *** mmaglana has quit IRC | 16:10 | |
| *** jistr has quit IRC | 16:10 | |
| joa | jeblair: seems it was one of the few links I didnt see :o | 16:11 |
| *** ihrachyshka_ has quit IRC | 16:12 | |
| joa | Well, I already have a basic CI working on sandbox. I'm merely trying to add the required definitions for the cinder-specific stuff (picking from infra/config) | 16:12 |
| jeblair | joa: oops :) anyway, you might want to try improving that instead of starting a new section in infra-manual. at least for starters, i'd like infra-manual to be the user manual for the system. maybe it can grow to be the admin manual too, but most of that material is on ci.openstack.org right now, and i don't want to bite of more than we can chew | 16:12 |
| jeblair | s/of/off/ | 16:13 |
| *** signed8bit is now known as signed8bit_ZZZzz | 16:14 | |
| joa | jeblair: well, what I started is only on the third_party.rst; which I discusses with anteaya and krtaylor. | 16:14 |
| joa | discussed* | 16:14 |
| *** signed8bit_ZZZzz has quit IRC | 16:14 | |
| joa | I really do feel that the links aren't easy to find for someone relatively new to openstack | 16:15 |
| jeblair | joa: right -- it's the first i've heard of it, and i want to help you do the work in the best place possible. i think that's in the doc i pointed out | 16:15 |
| joa | the goal was to try to have a single point to direct people trying to setup CI for 3rd-party testing to a single point | 16:15 |
| jeblair | joa: everything about our system is documented on ci.openstack.org | 16:15 |
| jeblair | joa: if something isn't directing people there, we should fix that | 16:16 |
| joa | well, it might be partly my fault oo | 16:16 |
| joa | I tend to follow the main page; and omit side-links or small links not indicated as important | 16:16 |
| jeblair | well, we can certainly re-org the docs -- i just want to avoid duplication :) | 16:17 |
| *** thedodd has quit IRC | 16:17 | |
| joa | I totally understand/agree | 16:17 |
| fungi | running-your-own.rst and third-party.rst are probably somewhat overlapping as it is, and both could use updating. perhaps the result could make it a little clearer which parts are needed to duplicate all of openstack infra and which bits are recommended for a generic third-party ci rig | 16:18 |
| jeblair | fungi: that would probably make a good change, yeah | 16:18 |
| joa | afaic see, runnin-your-own is a really "high level" howto, right ? | 16:19 |
| *** thedodd has joined #openstack-infra | 16:19 | |
| jeblair | joa: well, it was meant to be a pretty specific recipe for getting all the components up and running. i have not followed it myself. | 16:19 |
| joa | explains how to set things up with all the puppettery (I'm a total newbie in that) from infra/config; but I don't seem to see tips about how to pick the right bits for a "reduced" setup | 16:20 |
| asselin | running your own is not for 3rd party ci, right? | 16:20 |
| joa | asselin: well, it is related. | 16:20 |
| joa | but a 3rd-party only wants a subset of all that | 16:20 |
| joa | (imho) | 16:20 |
| jeblair | notably, one should omit gerrit. :) | 16:21 |
| joa | :p | 16:21 |
| asselin | right...I remember seeing that page a while back, but when I saw gerrit, etc. figured it's not for my use case. | 16:22 |
| lxsli | fungi: re: https://review.openstack.org/#/c/110653/ you earlier suggested inheriting from openstack_project::single_node_slave but on reluctantly cracking open the Puppet doc, I see: | 16:22 |
| lxsli | Note: This version of Puppet does not support using parameterized classes for inheritable base classes. The base class must have no parameters. | 16:22 |
| joa | hence my remark about how to choose which specific bits to use. | 16:22 |
| lxsli | http://docs.puppetlabs.com/puppet/latest/reference/lang_classes.html | 16:22 |
| *** atiwari has joined #openstack-infra | 16:22 | |
| lxsli | fungi: in this case would it be OK to duplicate single_node_slave.pp (and make my change) | 16:23 |
| fungi | lxsli: hopefully i didn't say "inherit" (i may have said "sub-class" but also might have misused the term since i don't know exactly how puppet refers to some concepts) | 16:23 |
| joa | jeblair: if you want to look/comment my patch on third_party.rst, if only to tell what's duplicating in your opinion, it will be a great help https://review.openstack.org/#/c/110651/ | 16:23 |
| fungi | lxsli: i'm looking over the change now and there are some shell script concerns, but i'm mostly trying to figure out how to unravel what you understand from a puppet perspective so i can provide clearer recommendations on how to adjust that | 16:24 |
| joa | as the commit msg says, it's a rough outline using my written feedback, but I do think it's worth it | 16:24 |
| lxsli | fungi: basically nothing sorry, I know a little Chef :/ | 16:24 |
| *** jgallard has quit IRC | 16:24 | |
| lxsli | Thanks for taking the time, I appreciate it | 16:25 |
| arosen | asselin: I'm starting fresh today with your github repo here: https://github.com/rasselin/os-ext-testing :) | 16:27 |
| *** shashankhegde has quit IRC | 16:27 | |
| joa | oh, a clone of jaypipes' | 16:27 |
| asselin | arosen, ok. just a warning, the os-ext-testing version is not up-to-date. i'll have to sync it up. | 16:28 |
| asselin | s/os-ext-testing/s-ext-testing-data/ | 16:28 |
| arosen | asselin as in your repo isn't up to date on github? | 16:28 |
| asselin | mainly there's extra entries needed to be added to vars.sh | 16:28 |
| *** derekh_ has quit IRC | 16:28 | |
| *** bdpayne has quit IRC | 16:29 | |
| *** jaypipes has joined #openstack-infra | 16:29 | |
| arosen | asselin: yea i noticed that. I just left the other entries with the values you had there for now. | 16:29 |
| asselin | sorry, let me say it again: the os-ext-testing and os-ext-testing-data (both my forks) are not in sync now. | 16:29 |
| asselin | I need to push up the latest to the os-ext-testing-data repo. Mainly the vars.sh needs updating. | 16:29 |
| *** hashar has quit IRC | 16:29 | |
| ociuhandu | jeblair: Hi. We have seen that the “recheck *” syntax is being declared as off-limits for 3rd party ci’s and there was a suggestion on changing it to “CI Name: recheck”. Can you please confirm that this is the way to go and by when does this have to be implemented? | 16:29 |
| arosen | asselin: are you planning on doing that today :) ? | 16:30 |
| asselin | arosen, yes, I'll plan to do that within later today. | 16:30 |
| jeblair | ociuhandu: http://ci.openstack.org/third_party.html says "Support the following syntaxes recheck no bug and recheck bug ###." | 16:31 |
| ociuhandu | jeblair: yes, that is for all CIs, I was asking for a trigger for one CI only, sorry for not being more clear in the first line | 16:33 |
| *** dmellado has quit IRC | 16:34 | |
| *** SpamapS_ is now known as SpamapS | 16:34 | |
| *** SpamapS has quit IRC | 16:34 | |
| *** SpamapS has joined #openstack-infra | 16:34 | |
| *** mmaglana has joined #openstack-infra | 16:34 | |
| *** marcoemorais has quit IRC | 16:34 | |
| jeblair | ociuhandu: i don't think a policy change regarding that has been finalized | 16:35 |
| *** marcoemorais has joined #openstack-infra | 16:35 | |
| *** pblaho has quit IRC | 16:35 | |
| *** flaper87 is now known as flaper87|afk | 16:35 | |
| *** marcoemorais has quit IRC | 16:35 | |
| arosen | asselin: Could not find class snmpd for ubuntu1204server.nicira.eng.vmware.com at /home/nicira/os-ext-testing/puppet/modules/os_ext_testing/manifests/base.pp:20 on node ubuntu1204server.nicira.eng.vmware.com Is there away to easily about setting up snmpd for now or using a gmail address there? | 16:35 |
| *** marcoemorais has joined #openstack-infra | 16:35 | |
| *** marcoemorais has quit IRC | 16:35 | |
| ociuhandu | jeblair: we currently use recheck hyper-v and from an email thread on the dev list I saw an email from sdague quoting “Consider: ^(recheck|check|reverify) off limits namespace.” | 16:35 |
| asselin | arosen, I didn't setup/test any snmp settings | 16:36 |
| ociuhandu | jeblair: so it is safe to assume that for the time being we can maintain the current syntax? | 16:36 |
| *** marcoemorais has joined #openstack-infra | 16:36 | |
| *** marcoemorais1 has joined #openstack-infra | 16:36 | |
| arosen | asselin: bash install_master.sh is stopping with that message. | 16:36 |
| asselin | arosen, i don't think it's necessary for 3rd party ci....??? | 16:36 |
| *** jgriffit1 has joined #openstack-infra | 16:37 | |
| *** marcoemorais1 has quit IRC | 16:37 | |
| asselin | arosen, I never got that error...let me take a look.... | 16:37 |
| *** marcoemorais1 has joined #openstack-infra | 16:37 | |
| *** marcoemorais1 has quit IRC | 16:37 | |
| *** marcoemorais1 has joined #openstack-infra | 16:37 | |
| *** mmaglana has quit IRC | 16:38 | |
| asselin | arosen, try commenting out that line #include snmpd | 16:39 |
| arosen | asselin: Could not find class sudoers for ubuntu1204server.nicira.eng.vmware.com at /home/nicira/os-ext-testing/puppet/modules/os_ext_testing/manifests/base.pp:37 on node ubuntu1204server.nicira.eng.vmware.com | 16:39 |
| arosen | asselin: i didn't hit this last time using jays install_master.sh script yesterday. | 16:40 |
| *** marcoemorais has quit IRC | 16:40 | |
| asselin | arosen, hmm...seems something went wrong with the initial part of the install script that imports infra's puppet scripts. | 16:40 |
| *** gokrokve has joined #openstack-infra | 16:40 | |
| jeblair | ociuhandu: we have made no provision for ci-specific recheck commands, either previously or as we are now changing our syntax. i do not think they should be supported (i think it's silly to ask developers to learn a complicated pseudo-language to interact with ci systems) | 16:41 |
| asselin | arosen, delete your install_puppet.sh script and run the script again | 16:41 |
| asselin | arosen, ok I think I know what it is | 16:43 |
| *** jpich has quit IRC | 16:43 | |
| asselin | I changed the top part of the script so that the os-ext-testing and os-ext-testing-data are sibling folders instead of parent child.(could wrap my head around with the original method) | 16:44 |
| *** MarkAtwood has quit IRC | 16:45 | |
| *** rcarrill` has joined #openstack-infra | 16:45 | |
| arosen | asselin: yea I accidentally had them parent child and fixed that. Seems like it's got past that part now | 16:46 |
| ociuhandu | jeblair: so as long as we also publish the recheck syntax on any fail message, we can keep the existing one? | 16:46 |
| krtaylor | ociuhandu, I think you are referring to this proposal by sdague https://review.openstack.org/#/c/109565/ | 16:47 |
| *** rcarrillocruz has quit IRC | 16:47 | |
| asselin | arosen, ok good. | 16:47 |
| krtaylor | ociuhandu, until that is decided, it is still the existing policy | 16:47 |
| *** nkubota has joined #openstack-infra | 16:48 | |
| jeblair | ociuhandu: soon "recheck hyper-vi" will also trigger openstack's zuul | 16:48 |
| *** arnaud has joined #openstack-infra | 16:48 | |
| ociuhandu | krtaylor: yes, I am. I have seen it in the dev email thread. | 16:48 |
| *** rickerc has quit IRC | 16:49 | |
| *** luqas has quit IRC | 16:50 | |
| *** ramashri has quit IRC | 16:51 | |
| jeblair | (i do not consider that situation ideal, but i believe i've been pretty clear that we have, can, and will change zuul's recheck syntax at any time, and that third party systems should not produce their own grammar) | 16:51 |
| *** ildikov has joined #openstack-infra | 16:51 | |
| jeblair | i think if we want there to be a third-party ci grammar, than a formal proposal like sdague's is needed | 16:52 |
| *** bdpayne has joined #openstack-infra | 16:52 | |
| jeblair | also, now that we no longer have a need to record bug numbers, and now that zuul is more sophisticated about how it enqueues changes, we may be able to eliminate recheck altogether and take change submissions via a web interface to zuul | 16:54 |
| ociuhandu | jeblair: but if there’s no third-party ci grammar, we can never ensure that a change in zuul’s recheck will not impact the third-party ci’s or better said overlap with that. | 16:55 |
| *** Ryan_Lane has joined #openstack-infra | 16:55 | |
| arosen | asselin: quick question: http://codepad.org/oMlshhwG | 16:57 |
| jeblair | ociuhandu: yes, that's why third-party ci's should not invent their own grammar and support only "recheck". again, the current policy provides no guarantees and supporting anything else is not recommended. if we approve a change like sdague's then we will have a policy of supporting it, and we will be more careful about making changes. | 16:57 |
| *** MarkAtwood has joined #openstack-infra | 16:57 | |
| asselin | arosen, grep for "check-tempest-dsvm-cinder-cert-hp-3par-iscsi" and delete it. This should be in the 'private' repo..I'll have to fix that | 16:57 |
| *** jcoufal has quit IRC | 16:58 | |
| *** marcoemorais1 has quit IRC | 17:00 | |
| *** marcoemorais has joined #openstack-infra | 17:01 | |
| *** dkranz has quit IRC | 17:01 | |
| *** Guest53224 has quit IRC | 17:02 | |
| *** marcoemorais has quit IRC | 17:02 | |
| *** Hal has joined #openstack-infra | 17:02 | |
| *** Hal is now known as Guest31978 | 17:03 | |
| *** marcoemorais has joined #openstack-infra | 17:03 | |
| *** shashankhegde has joined #openstack-infra | 17:04 | |
| *** nkubota has quit IRC | 17:04 | |
| *** leopoldj has quit IRC | 17:06 | |
| *** signed8bit has joined #openstack-infra | 17:06 | |
| *** signed8bit has quit IRC | 17:07 | |
| *** jerryz_ has joined #openstack-infra | 17:07 | |
| *** signed8bit has joined #openstack-infra | 17:07 | |
| *** morganfainberg_Z is now known as morganfainberg | 17:09 | |
| kgiusti | hi all - can anyone give me advice on how best to solve: http://lists.openstack.org/pipermail/openstack-dev/2014-July/041503.html | 17:10 |
| jerryz | jerryz_: testing | 17:10 |
| kgiusti | it's been suggested to change devstack, rather than config, to include the new platform specific packages. | 17:10 |
| openstackgerrit | William Van Hevelingen proposed a change to openstack-infra/config: Remove cleanup code for some module transitions https://review.openstack.org/110721 | 17:11 |
| kgiusti | would that allow the CI tests for olso.messaging pass? | 17:11 |
| *** chuckC has quit IRC | 17:12 | |
| openstackgerrit | William Van Hevelingen proposed a change to openstack-infra/config: Remove cleanup code for some module transitions https://review.openstack.org/110721 | 17:12 |
| *** prad has joined #openstack-infra | 17:13 | |
| fungi | kgiusti: which ci tests for olso.messaging? devstack-based integration tests or unit tests? | 17:13 |
| *** pradk has joined #openstack-infra | 17:14 | |
| *** prad has quit IRC | 17:14 | |
| fungi | i'm guessing you mean new unit tests at https://review.openstack.org/#/c/75815/12/tests/test_amqp_driver.py | 17:14 |
| *** pradk is now known as prad | 17:14 | |
| kgiusti | fungi: thanks - ya, the new unit tests are of primary concern at this point. | 17:14 |
| kgiusti | fungi: mocking out the dependencies was deemed a no-go. | 17:14 |
| openstackgerrit | lokesh s proposed a change to openstack-infra/config: Adding windows disk-image builder project https://review.openstack.org/110562 | 17:15 |
| kgiusti | fungi: doesn't result in any real test coverage. The unit tests actually run the driver 'down to the wire' | 17:15 |
| fungi | kgiusti: what's the expected timeframe before the dependencies would be packaged and available in a long-term-support server platform (e.g. where actual operators/end-users would consider running it)? | 17:15 |
| *** jerryz_ has quit IRC | 17:16 | |
| fungi | kgiusti: i'd be worried about churn from trying to shoehorn in an early implementation which won't actually work on real servers for 2+ years, but maybe that concern is unfounded | 17:16 |
| *** gangil has quit IRC | 17:16 | |
| *** jerryz_ has joined #openstack-infra | 17:17 | |
| *** dkranz has joined #openstack-infra | 17:17 | |
| *** praneshp has joined #openstack-infra | 17:17 | |
| fungi | kgiusti: that's my usual cart-before-the-horse concern when devs are wanting to add support and testing for cool new feature x which isn't actually part of a mainstream server distro anywhere | 17:18 |
| reed | fungi, when you have time, could you please run the ATC script again to get the new folks to invite? | 17:19 |
| *** sarob_ has joined #openstack-infra | 17:19 | |
| *** jerryz_ has quit IRC | 17:20 | |
| kgiusti | fungi: valid concern. The packages have been in EPEL for awhile - not sure of the timeframe for adoption into RHEL proper. I might be able to find someone who can comment more authoritatively. Debian is probably further out, though one of the qpid folks has submitted the packages via mentor, and has a sponsor. | 17:20 |
| arosen | asselin: btw I'm hitting that same issue i was yesterday in with your repo when i try to rerun: install_master.sh, "Could not parse configuration file: Certificate names must be lower case; see #1168" | 17:20 |
| kgiusti | fungi: beyond those distros, I can't comment. I'm an apache qpid dev. | 17:21 |
| *** sarob_ has quit IRC | 17:21 | |
| *** rkukura has quit IRC | 17:22 | |
| fungi | kgiusti: so for rhel it's probably solved since lots of operators run that with epel packages. ubuntu is probably a ways out given that they just released an lts and generally only do that every two years, but then again getting it into ubuntu cloud archive might not be too hard | 17:22 |
| *** jerryz_ has joined #openstack-infra | 17:23 | |
| fungi | kgiusti: my guess is that the ubuntu openstack team would probably add it to uca fairly shortly after it gets into debian/testing (which could be fast or slow depending on when the upload gets sponsored and whether that lands during the jessie release freeze) | 17:23 |
| kgiusti | fungi: for ubuntu the apache qpid devs have set up a ppa. | 17:23 |
| kgiusti | fungi: which would be kept up to date by qpid apache itself. | 17:24 |
| *** yamahata_ has joined #openstack-infra | 17:24 | |
| *** nkubota has joined #openstack-infra | 17:24 | |
| fungi | kgiusti: so far we've shied away from adding ppas for ubuntu with the exception of those backporting major components (so far only python interpreters) which are already present in later ubuntu releases and for which the same ubuntu package maintainers are also caring after those ppas | 17:25 |
| *** arnaud has quit IRC | 17:25 | |
| kgiusti | fungi: but the long term ubuntu goal for qpid is to have proton included in the distros, just like the qpidd stuff. | 17:25 |
| openstackgerrit | Ricardo Carrillo Cruz proposed a change to openstack-infra/config: Refactor the various artifact upload scripts into one https://review.openstack.org/109816 | 17:26 |
| fungi | kgiusti: but we've been trying off and on to incorporate ubuntu cloud archive on our test servers, which has a much shorter inclusion turn-around for openstack-specific dependencies than ubuntu itself at least | 17:26 |
| *** dkranz has quit IRC | 17:26 | |
| *** jerryz_ has quit IRC | 17:28 | |
| openstackgerrit | Ricardo Carrillo Cruz proposed a change to openstack-infra/config: Refactor the various artifact upload scripts into one https://review.openstack.org/109816 | 17:28 |
| *** medieval1 has quit IRC | 17:29 | |
| *** medieval1 has joined #openstack-infra | 17:29 | |
| openstackgerrit | Ricardo Carrillo Cruz proposed a change to openstack-infra/config: Refactor the various artifact upload scripts into one https://review.openstack.org/109816 | 17:30 |
| *** david-lyle has quit IRC | 17:30 | |
| *** david-lyle has joined #openstack-infra | 17:30 | |
| openstackgerrit | A change was merged to openstack-infra/config: Publish *-specs better https://review.openstack.org/110682 | 17:30 |
| *** svarnau has joined #openstack-infra | 17:30 | |
| *** gondoi is now known as zz_gondoi | 17:31 | |
| kgiusti | fungi: So, if we can't get the packages onto the CI systems for now, is there a 'plan b' that will provide enough test coverage of the new driver without having to force these packages onto the CI systems? | 17:32 |
| kgiusti | fungi: like some way to make those unit tests optional? | 17:33 |
| phschwartz | fungi: All clear | 17:33 |
| kgiusti | fungi: and perhaps setting up a 3rd party CI just to test that driver? | 17:33 |
| *** Sukhdev has quit IRC | 17:34 | |
| *** markwash has joined #openstack-infra | 17:34 | |
| *** salv-orlando has quit IRC | 17:34 | |
| *** markwash has quit IRC | 17:34 | |
| fungi | kgiusti: possibly they could be excluded from the upstream ci test runs (though presumably someone still needs to run them and report on their efficacy, preferably on proposed changes before approval), or oslo.messaging could grow a functional test suite which stands it up and adds whatever additional packages are needed (though pre-caching those on the devstack slaves could get a little interesting) | 17:35 |
| rcarrill` | fungi: hey jeremy, just pushed https://review.openstack.org/#/c/109816/11 . As you can see it now handles only jenkins/java packages publish | 17:35 |
| rcarrill` | pulled stuff from maven jjb shell blocks and put it there | 17:35 |
| *** rcarrill` is now known as rcarrillocruz | 17:35 | |
| *** david-lyle has quit IRC | 17:35 | |
| fungi | rcarrillocruz: awesome. i'll take a look | 17:35 |
| *** david-lyle has joined #openstack-infra | 17:36 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/infra-specs: Add config repo split spec https://review.openstack.org/110730 | 17:36 |
| fungi | kgiusti: though i'm just one opinion. other infra team reviewers very well may have other notions as to what's acceptable, or recommendations on alternative solutions i'm missing | 17:36 |
| rcarrillocruz | i have not added the stash (build plus push to tarballs.o.o) on the same change | 17:37 |
| rcarrillocruz | i will create another change for that | 17:37 |
| *** Ajaeger1 has joined #openstack-infra | 17:37 | |
| fungi | rcarrillocruz: the push to tarballs.o.o probably doesn't need to be extracted into a script since we're using simple publishers for that pretty much everywhere anyway | 17:38 |
| kgiusti | fungi: thanks - your opinion is appreciated! I'll ping the oslo.messaging devs with what I've learned, thanks! | 17:38 |
| *** chandankumar_ has joined #openstack-infra | 17:38 | |
| *** nkubota has quit IRC | 17:39 | |
| rcarrillocruz | fungi: i'd like to have some consistency, for some stuff the upload is via a publish, for others is via a builder script :S | 17:40 |
| *** jgriffit1 has quit IRC | 17:40 | |
| fungi | kgiusti: you're welcome. and i do think that regardless of how we end up working around it for the short term, your best bet is to focus on getting it sponsored into debian, imported into ubuntu's development train for their next short-cycle release, and into ubuntu cloud archive for the existing 14.04 lts release | 17:40 |
| *** harlowja_away is now known as harlowja | 17:40 | |
| *** rkukura has joined #openstack-infra | 17:40 | |
| jeblair | fungi, kgiusti: one of the goals of our distro support policy is to keep things simple and accessible for devs and consistent for operators; so getting it into a distro channel like epel (done!) or uca is the best bet. | 17:41 |
| fungi | jeblair: kgiusti: that does bring up one possibility... we could run an oslo.messaging unit test job on centos 7 (once we have images) which installs the package from epel and exercises those specific tests | 17:42 |
| *** mmaglana has joined #openstack-infra | 17:42 | |
| fungi | in fact, if it works on centos 6 under python 2.6, we might even be able to do it that way today | 17:43 |
| *** Guest17150 is now known as dansmith | 17:43 | |
| *** packet has joined #openstack-infra | 17:44 | |
| *** wenlock has quit IRC | 17:44 | |
| jeblair | i'm not super keen on running the unit tests twice, so, thinking ahead, we may only want to do that if we're prepared to say that centos7 is the preferred testing platform for py27 because it has faster access to new packages. ;) | 17:45 |
| openstackgerrit | Ivan Kolodyazhny proposed a change to openstack-infra/config: Make 'gate-cinder-pylint' job vote https://review.openstack.org/110734 | 17:45 |
| jeblair | i'm not dead set on that; just throwing out (slightly tongue in cheek) food for thought | 17:45 |
| *** james_li has quit IRC | 17:45 | |
| fungi | jeblair: reasonable. though exercising it only on the 2.6 unit tests for now could still be a viable path | 17:46 |
| *** rainya has quit IRC | 17:46 | |
| fungi | we've taken that tack for some things which only centos 6 had recent enough versions of (i think mongodb is the main example there?) | 17:46 |
| *** rainya has joined #openstack-infra | 17:47 | |
| *** mwagner_lap has quit IRC | 17:48 | |
| fungi | but this does highlight that we probably wand ubuntu cloud archive to factor into our testing on trusty early enough that we can catch incremental breakages rather than with precise where we were trying to jump light years ahead on versions of things in uca | 17:48 |
| fungi | s/wand/want/ | 17:48 |
| *** Longgeek has joined #openstack-infra | 17:48 | |
| Ajaeger1 | wow, a job for 76 hours in the check queue? That's impressive... | 17:49 |
| *** rainya has quit IRC | 17:49 | |
| Ajaeger1 | ah, I see fungi's comments... | 17:50 |
| fungi | Ajaeger1: it was running on jenkins05 when it fell offline over the weekend due to a host server issue in rackspace | 17:50 |
| fungi | Ajaeger1: it'll be there until zuul is restarted or the change gets a new patchset | 17:50 |
| *** _nadya_ has joined #openstack-infra | 17:50 | |
| Ajaeger1 | fungi: yeah, that explains it... | 17:50 |
| Ajaeger1 | shall I just send a new revision? ;) | 17:50 |
| fungi | Ajaeger1: if you care about it that much, i won't stop you ;) | 17:51 |
| Ajaeger1 | It just confuses me ;) | 17:51 |
| *** morganfainberg is now known as morganfainberg_Z | 17:51 | |
| fungi | a trivial rebase would probably preserve any existing code review votes if it has them | 17:51 |
| *** Longgeek_ has quit IRC | 17:51 | |
| fungi | since there were only a couple of changes still affected when i found it on monday, i just left comments on them explaining the situation and how to work around it | 17:52 |
| *** jedimike has quit IRC | 17:54 | |
| Ajaeger1 | trivial rebase did not really help - but it had no comments ;) | 17:54 |
| Ajaeger1 | fungi: but it depends on another one that just got rebased as well - again no votes | 17:54 |
| fungi | ahh | 17:55 |
| Ajaeger1 | (rebased as part of my rebase) | 17:55 |
| fungi | well, anyway, it got it unstuck now | 17:55 |
| openstackgerrit | Joe Gordon proposed a change to openstack-infra/elastic-recheck: Add fingerprint for bug 1350466 https://review.openstack.org/110737 | 17:55 |
| uvirtbot | Launchpad bug 1350466 in nova "deadlock in scheduler expire reservation periodic task" [Undecided,New] https://launchpad.net/bugs/1350466 | 17:55 |
| *** jedimike has joined #openstack-infra | 17:56 | |
| *** _nadya_ has quit IRC | 17:56 | |
| * Ajaeger1 has prepared a docs-specs patch to see whether publishing now works properly. Fungi, stevemar I'll keep you updated | 17:56 | |
| fungi | Ajaeger1: you may want to wait a little longer... i think jjb isn't done updating the jobs from the last template tweak | 17:58 |
| fungi | at least picking one at random it hasn't gotten adjusted yet | 17:59 |
| fungi | still has copy after failure on and the old source/destination values for the moment | 17:59 |
| *** rainya has joined #openstack-infra | 18:00 | |
| *** packet has quit IRC | 18:01 | |
| openstackgerrit | Michael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Updated project list interface. https://review.openstack.org/110169 | 18:01 |
| *** Longgeek has quit IRC | 18:01 | |
| arosen | I've got a question about http://ci.openstack.org/nodepool.html Why do we need to build new images? Can't we just always use the base ubuntu image? | 18:02 |
| Ajaeger1 | fungi, ok will wait some more and then check. Thanks | 18:02 |
| *** dustins has quit IRC | 18:03 | |
| fungi | arosen: we're pre-caching repetitive changes which would otherwise happen once the nodes are spun up or when jobs start | 18:03 |
| fungi | arosen: that maximizes efficiency and minimizes job turn-around | 18:04 |
| arosen | fungi: what kind of repetitive changes? | 18:04 |
| *** gyee has joined #openstack-infra | 18:04 | |
| arosen | like apt-get install mysql/git etc? | 18:04 |
| fungi | arosen: package updates, cloning git repos, et cetera | 18:04 |
| arosen | fungi: ah and then just git pull origin master on them so you don't need to download the whole thing each time? | 18:04 |
| *** mfer has quit IRC | 18:04 | |
| *** dustins has joined #openstack-infra | 18:04 | |
| fungi | yep | 18:04 |
| *** ZZelle_ has joined #openstack-infra | 18:04 | |
| fungi | we also pre-cache packages devstack may want to install, cirros/whatever images it wants to load into its glance, et cetera | 18:05 |
| fungi | so it not only saves time but also network utilization | 18:05 |
| *** gyee has quit IRC | 18:05 | |
| arosen | fungi: gotcha, trying to sort my way through getting ci going. I guess it's something i should also deploy :) | 18:05 |
| *** gyee has joined #openstack-infra | 18:05 | |
| fungi | and in fact that's one of our biggest gains. minimizing the extent to which we touch the provider network and internet substantially increases test reliability | 18:06 |
| *** packet has joined #openstack-infra | 18:06 | |
| Ajaeger1 | fungi: jenkins1,2,3,4,5 are fine for the one job that I checked... | 18:06 |
| fungi | Ajaeger1: https://jenkins01.openstack.org/job/gate-keystone-specs-docs/configure is the example i was looking at, and it still seems to use the old pattern | 18:07 |
| stevemar | AJaeger, can you link me? | 18:07 |
| fungi | oh, wait, that's the docs-draft publisher i'm looking at | 18:07 |
| Ajaeger1 | fungi: https://jenkins05.openstack.org/view/All/job/barbican-specs-publish-specs/configure | 18:07 |
| Ajaeger1 | fungi: indeed, that's the gate, not the publisher | 18:08 |
| Ajaeger1 | the gate hasn't changed ;) | 18:08 |
| fungi | Ajaeger1: yep, they look fine now--go for it | 18:08 |
| Ajaeger1 | ;) | 18:09 |
| *** juzuluag has quit IRC | 18:10 | |
| arosen | fungi: nodepool uploads the images to glance? | 18:14 |
| *** arnaud has joined #openstack-infra | 18:15 | |
| jesusaurus | arosen: currently it just does a nova snapshot-image | 18:17 |
| arosen | jesusaurus: cool that makes sense :) | 18:17 |
| *** gyee has quit IRC | 18:18 | |
| mordred | arosen: it's going to be doing glance uploads soon-ish | 18:18 |
| *** rcarrill` has joined #openstack-infra | 18:19 | |
| *** Ryan_Lane has quit IRC | 18:19 | |
| *** Ryan_Lane1 has joined #openstack-infra | 18:19 | |
| *** gyee has joined #openstack-infra | 18:19 | |
| *** jcoufal has joined #openstack-infra | 18:20 | |
| *** rcarrillocruz has quit IRC | 18:20 | |
| arosen | mordred: i guess that works as long as you're running on a cloud that allows you to do that. | 18:20 |
| *** banix has quit IRC | 18:20 | |
| arosen | I'm weary of the nova snapshot-image as i feel like it's likely another point of failure but I'm curious to see how it goes. | 18:21 |
| *** zns has quit IRC | 18:22 | |
| jesusaurus | ive been using snapshot-image for various things for a couple years. it can be slow, but it very rarely fails outright for me | 18:22 |
| *** laxity_ has quit IRC | 18:23 | |
| jeblair | jesusaurus: fails quite often for us :) | 18:23 |
| jeblair | jesusaurus: and also sometimes takes more than 8 hours. | 18:24 |
| jesusaurus | jeblair: the snapshotting? i almost never see it fail | 18:24 |
| jesusaurus | but yeah, sometimes it can be slooooowwwww | 18:24 |
| *** wenlock has joined #openstack-infra | 18:24 | |
| *** laxity_ has joined #openstack-infra | 18:24 | |
| *** hashar has joined #openstack-infra | 18:24 | |
| *** Guest42008 has quit IRC | 18:27 | |
| *** jp_at_hp has quit IRC | 18:28 | |
| *** hashar has quit IRC | 18:29 | |
| * Ajaeger1 throws a round of beer | 18:29 | |
| Ajaeger1 | fungi, stevemar : http://specs.openstack.org/openstack/docs-specs/ | 18:29 |
| stevemar | AJaeger, niiiice | 18:29 |
| jeblair | Ajaeger1: neato! i'm guessing we don't actually want the robots.txt though | 18:30 |
| Ajaeger1 | jeblair: see stevemar's patch https://review.openstack.org/110689 | 18:30 |
| Ajaeger1 | and here's the patch for the remaining repos: https://review.openstack.org/109675 | 18:31 |
| jeblair | stevemar: approved | 18:31 |
| stevemar | thx jeblair | 18:31 |
| Ajaeger1 | jeblair, fungi, stevemar: How do we great an index page in the best way? | 18:31 |
| *** otherwiseguy has quit IRC | 18:31 | |
| stevemar | i'm going to push another update to keystone-specs | 18:32 |
| *** kgiusti has quit IRC | 18:33 | |
| *** chandankumar_ has quit IRC | 18:33 | |
| *** _nadya_ has joined #openstack-infra | 18:34 | |
| *** lukego has joined #openstack-infra | 18:34 | |
| *** kgiusti has joined #openstack-infra | 18:34 | |
| Ajaeger1 | jeblair: I've put some information about publishing together at https://etherpad.openstack.org/p/doc_publishing | 18:34 |
| Ajaeger1 | Publishing of the books shouldn't be a large problem, I can enhance our build job to output which directory was changed | 18:35 |
| Ajaeger1 | Publishing docs.openstack.org/index.html and other files might need some special treatment... | 18:35 |
| jeblair | Ajaeger1: cool, i'm writing the first draft of the docs publishing spec now | 18:36 |
| Ajaeger1 | But we can use a separate publishing job that does not delete. | 18:36 |
| Ajaeger1 | jeblair: great! | 18:36 |
| fungi | Ajaeger1: there are a few possibilities... | 18:37 |
| openstackgerrit | A change was merged to openstack-infra/config: Remove the robots entry from specs.openstack.org https://review.openstack.org/110689 | 18:37 |
| *** marcoemorais has quit IRC | 18:37 | |
| *** mpaolino has quit IRC | 18:38 | |
| Ajaeger1 | fungi: I don't consider it a shipstopper, if we can handle 99 % of the files like I see it already coming together. If these few possibilities work out, it would be super-perfect ;) | 18:38 |
| *** marcoemorais has joined #openstack-infra | 18:38 | |
| *** marcoemorais has quit IRC | 18:38 | |
| Ajaeger1 | fungi: or are you speaking about the index generation for specs.o.o? | 18:38 |
| fungi | Ajaeger1: yeah | 18:38 |
| *** e0ne has quit IRC | 18:39 | |
| *** marcoemorais has joined #openstack-infra | 18:39 | |
| fungi | Ajaeger1: so, we could just puppet a static index into place | 18:39 |
| Ajaeger1 | fungi: yeah, that's the easiest part... | 18:39 |
| fungi | Ajaeger1: or have a job dynamically generate one and run as a dependent job to specs publishing jobs | 18:39 |
| *** marcoemorais has quit IRC | 18:39 | |
| *** marcoemorais has joined #openstack-infra | 18:40 | |
| fungi | Ajaeger1: or use a custom apache vhost template to style mod_autoindex with out own provided readme(header) and footer files | 18:40 |
| fungi | s/out/our/ | 18:40 |
| *** marcoemorais has quit IRC | 18:40 | |
| Ajaeger1 | IMO we need a rather static list - just a list of the form: OpenStack Documentation Program specs, Infra Program specs etc.... | 18:40 |
| *** marcoemorais has joined #openstack-infra | 18:40 | |
| Ajaeger1 | And I hope that the list of spec repositories stays static. | 18:40 |
| Ajaeger1 | Or do you want to list *all specs* instead of all *spec repositories*? In that case something dynamic is mandatory IMO | 18:41 |
| fungi | i'm kindof in favor of relying on mod_autoindex and just making it look more openstacky | 18:41 |
| jeblair | as long as it's just openstack, it shouldn't change too quickly (they are per-program, not per-project) | 18:41 |
| fungi | rather than reinventing page indexing, since apache already has a mechanism to give us one | 18:41 |
| jeblair | (i'm fine with autoindex or static) | 18:42 |
| fungi | i also am not *too* worried about the plain jane autoindex it has today. it may not be pretty but it's effective | 18:42 |
| *** marun has quit IRC | 18:42 | |
| Ajaeger1 | Yeah, the plain jane autoindex is there - just noticed;) | 18:43 |
| jeblair | regarding https://review.openstack.org/#/c/109675/5 i think that the 'specs-jobs' template should only include the -docs job since the others are not universal (yet) | 18:43 |
| Ajaeger1 | jeblair: the majority uses both gates, see https://review.openstack.org/106100 | 18:44 |
| *** Ryan_Lane1 has quit IRC | 18:44 | |
| *** james_li has joined #openstack-infra | 18:45 | |
| Ajaeger1 | jeblair: I can remove the py27 jobs again, if you think that's cleaner and update 109675. | 18:45 |
| Ajaeger1 | jeblair: but please first review how many jobs use the template already | 18:45 |
| *** zns has joined #openstack-infra | 18:47 | |
| *** Longgeek has joined #openstack-infra | 18:47 | |
| jeblair | Ajaeger1: yes, but it's not universal. this way all of the specs repos can use it, and then repos that have additional jobs can add them | 18:47 |
| arosen | jaypipes: i'm curious why here http://www.joinfu.com/2014/02/setting-up-an-openstack-external-testing-system-part-2/ or in general you need to setup a slave node and run your script. How does that node get cleaned up? | 18:47 |
| Ajaeger1 | jeblair: ok, let me rework 109765... | 18:48 |
| *** rcarrillocruz has joined #openstack-infra | 18:48 | |
| jaypipes | arosen: it doesn't. it's meant to be a single-use slave node... | 18:49 |
| jaypipes | arosen: nodepool (the work that asselin is adding to the os-ext-testing repo) takes care of keeping a pool of single-use slave nodes for running test jobs | 18:49 |
| arosen | jaypipes: ah i see. So i guess nodepool manages the life cycle of a pool of nodes for you i guess. | 18:50 |
| arosen | jaypipes: thanks :) | 18:50 |
| jaypipes | arosen: we found that trying to "clean" or "reset" the slave node after devvstack/tempest runs was just too unreliable compared to blowing away the node :) | 18:50 |
| *** rcarrill` has quit IRC | 18:50 | |
| arosen | jaypipes: yea makes sense, definitely hard to clean up after devstack | 18:50 |
| *** samy_ has joined #openstack-infra | 18:50 | |
| *** ociuhandu has quit IRC | 18:51 | |
| Ajaeger1 | fungi: do you have an example for a static html page we deploy via puppet? | 18:51 |
| fungi | Ajaeger1: plenty, but i'll find you a good clean example one | 18:52 |
| Ajaeger1 | thanks | 18:52 |
| *** e0ne has joined #openstack-infra | 18:52 | |
| *** jerryz_ has joined #openstack-infra | 18:53 | |
| fungi | Ajaeger1: http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/manifests/zuul_prod.pp#n111 | 18:53 |
| *** _nadya_ has quit IRC | 18:53 | |
| fungi | Ajaeger1: and the corresponding source file for that is stored as http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/files/zuul/scoreboard.html | 18:54 |
| Ajaeger1 | thanks, fungi | 18:54 |
| fungi | Ajaeger1: we can get slightly fancier by making that an erb template instead with an iterator in it, and then keeping an array in the puppet manifest, but that simple case is a good start | 18:55 |
| *** hashar has joined #openstack-infra | 18:55 | |
| Ajaeger1 | fungi: let me try starting plain simple and then somebody can make it really fancy ;) | 18:55 |
| fungi | Ajaeger1: good call ;) | 18:55 |
| arosen | nodepool needs to run as an admin user against an openstack deployment? | 18:56 |
| fungi | arosen: nope, just a regular old tenant user | 18:56 |
| fungi | arosen: needs to be able to list, create and delete instances, snapshot them and boot from snapshots, list images and snapshot images, et cetera | 18:57 |
| *** jerryz_ has quit IRC | 18:57 | |
| fungi | arosen: needs to be able to delete snapshot images too | 18:57 |
| fungi | arosen: the upstream use case is that we've got basically vanilla accounts with multiple public cloud providers and so we needed a way to manage our test instance life cycle in them with some automation | 18:58 |
| arosen | fungi: cool, that's how i prefer to run this against our cloud as well :) | 18:59 |
| fungi | arosen: depending on the provider, may also need to be able to create/associate/delete floating ips, select interfaces onto specific neutron networks, and so on | 19:00 |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish additional specs repositories https://review.openstack.org/109675 | 19:00 |
| *** rcarrill` has joined #openstack-infra | 19:00 | |
| Ajaeger1 | jeblair: I implemented your suggestion ^ | 19:00 |
| jeblair | Ajaeger1: cool, thanks! | 19:00 |
| arosen | fungi: so i sorta have nodepool running but it's complaining about: BadRequest: Multiple possible networks found, use a Network ID to be more specific. | 19:01 |
| *** gangil has joined #openstack-infra | 19:01 | |
| fungi | arosen: yep, so you've got neutron presenting multiple networks into your tenant | 19:01 |
| *** kgiusti has quit IRC | 19:01 | |
| *** rcarrillocruz has quit IRC | 19:01 | |
| arosen | fungi: one problem I think we'll have with nodepool is i'd prefer to not use dhcp as it can fail from time to time. | 19:01 |
| arosen | fungi: right now the way the vmware ci works is we create a new network and have a prebaked image with a static ip set in it. | 19:01 |
| arosen | fungi: I guess if i use this i have to rely on dhcp / | 19:02 |
| fungi | arosen: you can see how we use the net-id parameter in neutron-based providers here... http://git.openstack.org/cgit/openstack-infra/config/tree/modules/openstack_project/templates/nodepool/nodepool.yaml.erb | 19:02 |
| *** kgiusti has joined #openstack-infra | 19:03 | |
| fungi | arosen: i'm not sure it relies on dhcp specifically, it just relies on nova to choose a distinct ip address unrelated to the contents of the base/snapshot images | 19:03 |
| fungi | arosen: i think one of our providers is using file injection to set ip addresses, for example | 19:03 |
| arosen | fungi: right i mean it relies on the guest getting an ip address | 19:04 |
| *** zul has quit IRC | 19:04 | |
| *** cdub has quit IRC | 19:04 | |
| arosen | fungi: For now i'll just rely on dhcp and cross my fingers :). It also looks like it relies on metadata | 19:04 |
| arosen | for key keypair injection | 19:05 |
| fungi | arosen: that definitely sounds like a complication. vmware-backed nova in general has dhcp issues, or just your particular environment? | 19:05 |
| arosen | but we can rebake the key. | 19:05 |
| arosen | fungi: i'd say the neutron-dhcp-agent isn't very robust so from time to time we have failures related to that | 19:06 |
| arosen | so we usually just spawn each CI run it it's own network with the ip statically configured in the image. | 19:06 |
| fungi | arosen: one upside to nodepool in this context is that it will boot the instance from a snapshot, test that it's reachable, then add it to the available pool or delete it depending on whether it's working | 19:06 |
| arosen | fungi: does nodepool handing allocating floatingip's too ? | 19:07 |
| fungi | arosen: so some background number of instances not getting ip addresses may go entirely unnoticed as they'll just be automatically recycled into the quota again | 19:07 |
| *** e0ne has quit IRC | 19:07 | |
| fungi | arosen: it does, yes | 19:07 |
| fungi | arosen: we use fip on hpcloud, in fact | 19:07 |
| *** marun has joined #openstack-infra | 19:09 | |
| *** chuckC has joined #openstack-infra | 19:09 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Create specs.o.o/index.html https://review.openstack.org/110756 | 19:10 |
| *** garyh has quit IRC | 19:10 | |
| Ajaeger1 | fungi: Static and plain solution ^ ;) | 19:10 |
| Ajaeger1 | fungi: Does anybody need to remove the robots.txt from specs.o.o or is this done via puppet? | 19:11 |
| arosen | fungi: have you seen this error? http://paste.openstack.org/show/89220/ | 19:11 |
| *** nelsnelson has quit IRC | 19:11 | |
| fungi | Ajaeger1: i'll have to manually clean up the files in it for now | 19:12 |
| *** mestery has quit IRC | 19:12 | |
| arosen | fungi: oh maybe cause i have ' ' around net-id let me remove those. | 19:12 |
| fungi | Ajaeger1: same for the stray subdirs | 19:12 |
| arosen | fungi: nah same error. | 19:12 |
| *** garyh has joined #openstack-infra | 19:12 | |
| *** mestery has joined #openstack-infra | 19:12 | |
| Ajaeger1 | fungi: good that we only tested with keystone-specs, so little to remove. You can put it on your cleanup list now ;) Thanks! | 19:14 |
| *** mestery has quit IRC | 19:14 | |
| fungi | arosen: doesn't look familiar. are you sure the network list for that tenant shows that as a valid network uuid? | 19:14 |
| *** mestery has joined #openstack-infra | 19:14 | |
| arosen | fungi: gah sorry it was a typo I had -net-id but it needs to be - net-id X.x | 19:15 |
| fungi | arosen: oh, yep. like python, yaml is fairly picky about whitespace | 19:16 |
| fungi | nodepool could probably stand to grow a configuration validator at some point | 19:17 |
| fungi | beyond just the simple "does this parse as yaml" | 19:17 |
| *** jerryz_ has joined #openstack-infra | 19:18 | |
| *** _nadya_ has joined #openstack-infra | 19:19 | |
| fungi | Ajaeger1: cleaned up, though note that this is going to suffer from the same "files never go away" issue our current docs publishing has, so if projects move/delete specs from their repos, they'll hang around as cruft on the site | 19:19 |
| arosen | fungi: woot, nodepool just deployd an instance for me but it failed saying FloatingIpPoolNotFound. Do i need to precreate floatingips if i want to use them? | 19:20 |
| Ajaeger1 | fungi: ok, understood | 19:20 |
| Ajaeger1 | fungi: thanks! | 19:20 |
| fungi | Ajaeger1: so the sooner we get new docs publishing into place, the sooner this can take advantage of the same benefits | 19:20 |
| arosen | fungi: I had deployed it on a neutron network that is no-snat so it doesn't need a floatingip but i'm fine to go to a floatingip model to get things rolling. | 19:20 |
| fungi | arosen: i'm not sure. at the moment we use floating ips in the same providers where we use neutron networks, so it may be that the two are intrinsically linked from nodepool's perspective | 19:21 |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Publish additional specs repositories https://review.openstack.org/109675 | 19:22 |
| *** arnaud has quit IRC | 19:23 | |
| fungi | arosen: do you have the entire traceback from your FloatingIpPoolNotFound exception? looks like nodepool doesn't throw that directly so probably coming from novaclient | 19:24 |
| arosen | fungi: i do: http://codepad.org/HCCdxVfg | 19:25 |
| *** zz_gondoi is now known as gondoi | 19:25 | |
| fungi | ahh, so nodepool is trying to client.floating_ips.create() | 19:25 |
| jeblair | fungi: working on it! writing specs takes even longer than reviewing them! :) | 19:26 |
| arosen | fungi: i think this is actually a nova bug. This command works fine local env but not on what we have deployed | 19:26 |
| jeblair | especially if one avoids hand-waving | 19:26 |
| fungi | jeblair: yep, i hear that. just making sure we know that some of the specs publishing assumptions seem not to have taken cruft into account | 19:26 |
| fungi | arosen: oh, neat! | 19:27 |
| *** _nadya_ has quit IRC | 19:27 | |
| arosen | fungi: found it :), so you need to have this set in nova.conf on your nova-api servers | 19:27 |
| arosen | default_floating_pool = public | 19:27 |
| rcarrill` | fungi: if you have bandwidth to review https://review.openstack.org/#/c/109816/11 that'd be great, so i can look at it tomorrow in my morning | 19:28 |
| arosen | fungi: we should add support to nodepool to let us specify and go through neutron as we have multiple public networks | 19:28 |
| jeblair | fungi: yeah, i think that's the right choice; it's a common problem that we should fix all at once | 19:28 |
| rcarrill` | jeblair: https://review.openstack.org/#/c/109816/11 is the change for refactoring the jenkins/java packages upload to maven , we chat about this last week | 19:28 |
| arosen | let me see if this setting was in havana. If so i'll push it out to our cloud just picking one. | 19:28 |
| *** rcarrill` is now known as rcarrillocruz | 19:29 | |
| *** elambert has joined #openstack-infra | 19:30 | |
| *** dguerri is now known as _dguerri | 19:31 | |
| fungi | arosen: yeah, to some extent the neutron network provider support in nodepool grew out of people who were trying to use it in their own neutron-based clouds (lucky us, because then we ended up with one or two we needed to connect to after that), so if you have more efficient ways we could be interacting with it then we appreciate patches | 19:31 |
| fungi | rcarrillocruz: one suggestion for 109816, which is nto really a critique of the patch itself, is that it would be good to see the config change as a child of that change so it's more obvious where and how this script will start getting called | 19:33 |
| *** shashankhegde has quit IRC | 19:34 | |
| *** banix has joined #openstack-infra | 19:34 | |
| arosen | fungi: cool, i'm definitely happy to contribute some patches back to allow a way to better leverage neutron deployments. | 19:34 |
| *** SumitNaiksatam has quit IRC | 19:34 | |
| *** jerryz_ has quit IRC | 19:35 | |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for grenade quota reserve deadlock bug 1350064 https://review.openstack.org/110481 | 19:35 |
| uvirtbot | Launchpad bug 1350064 in nova "Deadlock in quota reservations in security groups tests on old side of grenade (icehouse)" [Undecided,New] https://launchpad.net/bugs/1350064 | 19:35 |
| rcarrillocruz | fungi: you mean how jjb jobs would call the script? | 19:36 |
| fungi | rcarrillocruz: right | 19:36 |
| fungi | rcarrillocruz: also, it looks like the latest patchset of 109816 stopped incorporating the pypi uploads... did they end up being too hard to combine with the others, or is that just slated as an incremental improvement later? | 19:37 |
| *** e0ne has joined #openstack-infra | 19:37 | |
| rcarrillocruz | fungi: i see as an incremental improvement later | 19:37 |
| *** packet has quit IRC | 19:37 | |
| fungi | rcarrillocruz: that's perfectly fair | 19:37 |
| fungi | rcarrillocruz: this way the change is more bite-sized | 19:38 |
| rcarrillocruz | fungi: java/jenkins uploads are treated as builder scripts right now, whereas pypi stuff are publishers | 19:38 |
| rcarrillocruz | fungi: yep, i prefer doing changes more self-contained, i find them easier for reviewers to review | 19:38 |
| fungi | rcarrillocruz: definitely. me too | 19:38 |
| fungi | rcarrillocruz: like i always say, review unto others as you would have them review unto you | 19:39 |
| rcarrillocruz | heh, inded | 19:39 |
| *** ramashri has joined #openstack-infra | 19:40 | |
| *** masayukig has quit IRC | 19:43 | |
| arosen | fungi: cool, looks like node pool is working fine. It launched my instances and is now running commands in them. | 19:43 |
| mordred | woot! | 19:44 |
| arosen | fungi: Do i need to do something now to configure jenkins to hook into nodepool so it can learn about my nodes? | 19:44 |
| mordred | arosen: there is a section of nodepool config where you tell it about your jenkins host(s) and username/api key and nodepool will add them as slaves | 19:44 |
| fungi | arosen: nodepool should be configured with credentials to connect to jenkins and register the slaves into them. also jenkins needs the zmq event publisher plugin so that nodepool can connect and learn about nodes which have started to get used or are done runnign jobs | 19:45 |
| *** e0ne has quit IRC | 19:45 | |
| mordred | arosen: https://github.com/openstack-infra/config/blob/master/modules/openstack_project/templates/nodepool/nodepool.yaml.erb#L832 for instance | 19:45 |
| *** masayukig has joined #openstack-infra | 19:48 | |
| openstackgerrit | David Lenwell proposed a change to openstack-infra/infra-manual: Changed width of the image at the top of the page https://review.openstack.org/108867 | 19:48 |
| *** mmaglana has quit IRC | 19:49 | |
| arosen | thanks, gotta figure out how to setup a user in jenkins to do this. | 19:51 |
| *** dustins has quit IRC | 19:52 | |
| *** james_li has quit IRC | 19:52 | |
| *** zns has quit IRC | 19:52 | |
| *** adalbas has quit IRC | 19:54 | |
| *** crc32 has joined #openstack-infra | 19:54 | |
| *** flaper87|afk is now known as flaper87 | 19:57 | |
| *** otherwiseguy has joined #openstack-infra | 19:57 | |
| *** james_li has joined #openstack-infra | 19:58 | |
| arosen | I just restarted nodepool since i made a config change to setup the jenkins stuff. It doesn't seem to really be doing anything now | 20:00 |
| arosen | it just says: DEBUG:nodepool.NodePool:No current image for d-p-c on local_01 | 20:00 |
| arosen | Should i manually cleanup the instance it had deployed? | 20:00 |
| *** arnaud has joined #openstack-infra | 20:02 | |
| *** arnaud__ has joined #openstack-infra | 20:02 | |
| fungi | arosen: you can 'sudo nodepool delete NNNN' where NNNN is the node id number | 20:02 |
| *** zns has joined #openstack-infra | 20:03 | |
| fungi | right now there are only a handful of nodepool cli subcommands, but the ones it has are pretty handy and you will want to familiarize yourself with them | 20:03 |
| Ajaeger1 | fungi: I guess I should call it a day ;) Thanks for the reviews, I'll do a final change and then sign off... | 20:04 |
| fungi | Ajaeger1: thanks as always@ | 20:04 |
| *** e0ne has joined #openstack-infra | 20:05 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Create specs.o.o/index.html https://review.openstack.org/110756 | 20:05 |
| *** yjiang5 has joined #openstack-infra | 20:06 | |
| *** e0ne has quit IRC | 20:06 | |
| *** Ajaeger1 has quit IRC | 20:07 | |
| *** shashankhegde has joined #openstack-infra | 20:07 | |
| yjiang5 | hi, all, I interrupt a git review w/ ctrl^C, then I can't push my patch anymore. Can anyone help me to have a look? The log is http://paste.openstack.org/show/89235/ | 20:08 |
| *** alexpilotti has quit IRC | 20:09 | |
| *** rcarrill` has joined #openstack-infra | 20:09 | |
| *** rcarrillocruz has quit IRC | 20:10 | |
| *** _dguerri is now known as dguerri | 20:10 | |
| fungi | yjiang5: what does 'git status' say? | 20:12 |
| *** stevemar has quit IRC | 20:12 | |
| yjiang5 | fungi: thanks for reply. git status is quite clean. | 20:13 |
| fungi | yjiang5: so it doesn't say you're in the middle of a rebase or anything... | 20:13 |
| yjiang5 | fungi: no | 20:13 |
| *** jamespage_ has joined #openstack-infra | 20:13 | |
| yjiang5 | fungi: http://paste.openstack.org/show/89236/ is the output of the git status. | 20:14 |
| fungi | yjiang5: judging from the error message, you've changed the gitsha on the parent commit unnecessarily... try 'git rebase 7c0ac50' and see if it's any better afterward | 20:15 |
| *** vhoward has left #openstack-infra | 20:15 | |
| yjiang5 | fungi: thanks. | 20:15 |
| openstackgerrit | Aishwarya Thangappa proposed a change to openstack-infra/storyboard: Adding Subscription Resource Hook https://review.openstack.org/102842 | 20:16 |
| yjiang5 | fungi: cool, it works !! | 20:17 |
| yjiang5 | fungi: thanks. | 20:18 |
| arosen | fungi: hrm any idea why it's looping only saying this: http://paste.openstack.org/show/89238/ ? | 20:19 |
| *** ihrachyshka has joined #openstack-infra | 20:19 | |
| fungi | arosen: it's saying it thinks it's configured to keep a minimum of two d-p-c nodes in a ready state but doesn't (yet) have a snapshot image built which corresponds to that label | 20:21 |
| fungi | arosen: hopefully, if things are going really well for you today, you'll see 'sudo nodepool image-list' say something about an image in a building state | 20:22 |
| *** rcarrillocruz has joined #openstack-infra | 20:22 | |
| fungi | arosen: and you can check the image.log for details | 20:22 |
| *** rcarrill` has quit IRC | 20:24 | |
| arosen | fungi: yup i had one in building and another in deleteing | 20:24 |
| *** e0ne has joined #openstack-infra | 20:25 | |
| arosen | but i think they were stuck since i restarted nodepool. I dod image-delete both of them and now a new one is building | 20:25 |
| fungi | arosen: that *probably* means building the snapshot is failing somewhere along the line, and you'll want to analyze the log to find out why | 20:25 |
| fungi | but yeah you might have just interrupted it mid-build earlier or something | 20:25 |
| arosen | Ah since i'm running nodepool in the forground there is no long | 20:25 |
| fungi | oh, right-o | 20:25 |
| arosen | but i think rebooting the process messed it up | 20:25 |
| fungi | yeah, it would | 20:26 |
| arosen | man this stuff is cool :) | 20:26 |
| *** jcoufal has quit IRC | 20:26 | |
| fungi | also hopefully when you're running nodepoold in the foreground you're doing it as whatever local user account you're intending to run the headless daemon under later, so it's not creating files it won't be able to read/modify/delete in production later | 20:27 |
| *** _nadya_ has joined #openstack-infra | 20:27 | |
| *** shashankhegde has quit IRC | 20:28 | |
| *** zns has quit IRC | 20:29 | |
| *** ihrachyshka has quit IRC | 20:31 | |
| *** ihrachyshka has joined #openstack-infra | 20:31 | |
| *** mmaglana has joined #openstack-infra | 20:32 | |
| arosen | fungi: yup I did sudo su zuul first :) | 20:32 |
| arosen | Doh getting this: | 20:32 |
| arosen | INFO:nodepool.image.build.local_01.d-p-c:Error: Non-HTTP proxy URI: | 20:33 |
| arosen | INFO:nodepool.image.build.local_01.d-p-c:Error: Try 'puppet help module install' for usage | 20:33 |
| *** andreaf has quit IRC | 20:34 | |
| fungi | arosen oof, that's fun | 20:35 |
| arosen | fungi: do you know what it means :) | 20:35 |
| rcarrillocruz | fungi: thx for the review, appreciate that, i will address suggestions tomorrow | 20:35 |
| fungi | arosen: puppet is having trouble retrieving something through a proxy which it doesn't understand? it's a little opaque to me | 20:36 |
| * fungi is stepping out for a little while but will return in an hour or so | 20:36 | |
| arosen | fungi: seems like things are continuing on so maybe it's okay | 20:37 |
| arosen | fungi: later. Huge thanks for your help! | 20:37 |
| *** zns has joined #openstack-infra | 20:38 | |
| *** jamespage_ has quit IRC | 20:39 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack-infra/config: Create specs.o.o/index.html https://review.openstack.org/110756 | 20:41 |
| *** andreykurilin has joined #openstack-infra | 20:44 | |
| *** unicell has quit IRC | 20:47 | |
| *** dkliban has quit IRC | 20:49 | |
| *** derekh_ has joined #openstack-infra | 20:49 | |
| reed | how excited am I about using storyboard? | 20:49 |
| krotscheck | reed: I dunno, are you? | 20:49 |
| reed | A: a *lot* | 20:49 |
| reed | krotscheck, I just discovered the possibility to create tasks within a story | 20:50 |
| reed | sooooooooooo cool! | 20:50 |
| krotscheck | Ahhh, that | 20:50 |
| krotscheck | That form needs a little love | 20:50 |
| *** MarkAtwood has quit IRC | 20:51 | |
| * krotscheck has been focused on getting search done. | 20:51 | |
| reed | it's already very good as is | 20:51 |
| krotscheck | Because we’re _so_ close on that one. | 20:51 |
| reed | i'm really excited | 20:51 |
| openstackgerrit | A change was merged to openstack-infra/elastic-recheck: Add query for grenade quota reserve deadlock bug 1350064 https://review.openstack.org/110481 | 20:51 |
| uvirtbot | Launchpad bug 1350064 in nova "Deadlock in quota reservations in security groups tests on old side of grenade (icehouse)" [Undecided,New] https://launchpad.net/bugs/1350064 | 20:51 |
| krotscheck | reed: We’re adding fulltext natural language search, and propagating the taggy UI to the type-specific lists. | 20:52 |
| reed | I'm working on a backlog for the groups portal, as a first exercise | 20:52 |
| *** doude has quit IRC | 20:52 | |
| *** wilfredo has joined #openstack-infra | 20:52 | |
| reed | krotscheck, I've asked Bitergia folks to start looking into it so that we can have reporting for it too, once the larger projects switch to sb | 20:52 |
| krotscheck | reed: Then you’ll be interested in this: https://review.openstack.org/#/c/105252/ | 20:53 |
| *** ociuhandu has joined #openstack-infra | 20:53 | |
| * reed checks | 20:53 | |
| krotscheck | reed: The “subscription” api is actually being driven off of a stream of event notifications on changes in the system. That spec exposes that stream so that third party tools can report off of a live stream of changes. | 20:54 |
| reed | neat | 20:54 |
| jeblair | reed: storyboard's API may change _dramatically_ in the near term; we're not ready to present a stable api to any consumers yet | 20:54 |
| reed | I think bicho will probably use the REST api anyway, it's an 'async' tool | 20:55 |
| jeblair | reed: so i mean, it's totally cool if you want to have the bitgeria folks take an early look... just no hand wringing when we change everything :) | 20:55 |
| *** elambert has quit IRC | 20:55 | |
| reed | jeblair, thanks for the heads up | 20:55 |
| *** hashar has quit IRC | 20:55 | |
| reed | jeblair, let me know when you're getting ready to switch infra to it so I can have bitergia ready by then | 20:56 |
| *** laxity_ has quit IRC | 20:56 | |
| reed | until then, I'll tell them not to rush anything | 20:56 |
| *** laxity_ has joined #openstack-infra | 20:56 | |
| jeblair | reed: ack | 20:57 |
| arosen | nodepool doesn't have a way to specify min-cpu? | 20:58 |
| arosen | i just grepped the code for cpu and nothing turned up. | 20:59 |
| *** zaneb has left #openstack-infra | 21:00 | |
| mordred | arosen: we configure ours by min ram - and then if there are more than one matching flavor (hi rackspace) we have a flavor name filter that can be used | 21:00 |
| arosen | mordred: ah cool, flavor name is what i was looking for.. We have to many random flavors with 16gigs 1 cpu etc lol | 21:01 |
| *** rlandy has quit IRC | 21:03 | |
| *** mbacchi has quit IRC | 21:05 | |
| openstackgerrit | Matt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for VirtualInterfaceCreateException https://review.openstack.org/110788 | 21:05 |
| *** julim has joined #openstack-infra | 21:07 | |
| *** zns has quit IRC | 21:08 | |
| *** shashankhegde has joined #openstack-infra | 21:09 | |
| *** weshay has quit IRC | 21:10 | |
| *** pcm_ has joined #openstack-infra | 21:12 | |
| *** Longgeek has quit IRC | 21:13 | |
| *** kgiusti has quit IRC | 21:14 | |
| *** jcoufal has joined #openstack-infra | 21:16 | |
| jeblair | arosen: i'd definitely be open to a change that made it so that only specifying the flavor name was supported | 21:17 |
| *** dkliban has joined #openstack-infra | 21:17 | |
| *** jedimike has quit IRC | 21:20 | |
| arosen | mordred: i don't see a flavor name filter in the code. | 21:21 |
| *** laxity_ has quit IRC | 21:22 | |
| arosen | Just got everything up and running. I had to move it to a network that didn't have any proxy stuff blocking anything. Fingers crossed. | 21:22 |
| *** laxity_ has joined #openstack-infra | 21:22 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/infra-specs: Add docs publishing spec https://review.openstack.org/110793 | 21:22 |
| jeblair | AJaeger: ^ | 21:23 |
| *** jcoufal has quit IRC | 21:23 | |
| *** emagana has joined #openstack-infra | 21:23 | |
| redrobot | sooooo.... any infra core devs here? I would appreciate it if somoene can move this CR along https://review.openstack.org/#/c/109398/ It's just a one-liner so it shouldn't take very long. | 21:23 |
| arosen | OH, so nodepool streams the output from the node it's building on? | 21:25 |
| arosen | That's why i'm still getting this proxy error i bet | 21:25 |
| arosen | actually nvm there isn't a proxy there. | 21:26 |
| arosen | anyone know what's causing this error? http://codepad.org/kBidlaS9 | 21:27 |
| *** kgiusti has joined #openstack-infra | 21:29 | |
| *** esker has quit IRC | 21:30 | |
| *** salv-orlando has joined #openstack-infra | 21:30 | |
| *** esker has joined #openstack-infra | 21:30 | |
| *** rwsu has quit IRC | 21:31 | |
| *** marun has quit IRC | 21:33 | |
| fungi | okay, back again | 21:34 |
| *** pcm__ has joined #openstack-infra | 21:34 | |
| *** esker has quit IRC | 21:35 | |
| *** Guest31978 has quit IRC | 21:36 | |
| *** julim has quit IRC | 21:37 | |
| *** pcm_ has quit IRC | 21:37 | |
| *** wilfredo has quit IRC | 21:39 | |
| *** rwsu has joined #openstack-infra | 21:39 | |
| arosen | fungi: trying to catch my vm and kill nodepool so i can debug this error :/ | 21:40 |
| arosen | is there a setting so it doesn't delete the instance? | 21:40 |
| fungi | arosen: hrm, i don't think there's a hold for image build template instances like there is for nodes themselves | 21:41 |
| fungi | arosen: i'd probably just kill nodepoold partway through the build and then get onto the template instance however you can manage | 21:41 |
| *** Ryan_Lane has joined #openstack-infra | 21:41 | |
| arosen | yup that's what i'm going to try and do . | 21:42 |
| arosen | I was to slow last time. | 21:42 |
| fungi | arosen: are you perchance specifying http_proxy_host and/or http_proxy_port anywhere in your puppet configuration? | 21:42 |
| *** kgiusti has quit IRC | 21:42 | |
| arosen | this is actually occuring on the guest i'm booting i think. | 21:42 |
| fungi | arosen: right, it's happening where nodepoold has ssh'd into the new template instance and is truing to install puppet modules it looks like | 21:43 |
| arosen | i don't set http_proxy anywhere that i can tell. | 21:43 |
| fungi | arosen: is it possible there are proxy-related envvars being set in the shell environment on that new instance, maybe globally in /etc/environment or something? | 21:44 |
| fungi | arosen: might be something baked into your base image | 21:44 |
| *** cdub has joined #openstack-infra | 21:45 | |
| arosen | fungi: yea that's what i'm thinking :(. I know there is a proxy setting baked into the image in /etc/apt.d but it looks like nodepool is smart enough to remove that. | 21:45 |
| arosen | /etc/apt/apt.conf | 21:45 |
| fungi | yeah, i don't think /etc/apt/apt.conf will have any bearing on whatever backend puppet uses to retrieve modules | 21:46 |
| asselin | arosen, did you adjust the proxy settings in vars.sh? | 21:48 |
| arosen | asselin: I had comment all of those settings out in vars. | 21:48 |
| fungi | arosen: according to http://docs.puppetlabs.com/puppet/latest/reference/modules_installing.html it tries to respect the http_proxy and https_proxy envvars in the calling environment | 21:48 |
| arosen | we'll see i'm almost to the pint where I can kill node-pool | 21:48 |
| *** markmcclain has quit IRC | 21:49 | |
| *** bauzas has quit IRC | 21:50 | |
| asselin | arosen, I spent a lot of time debugging proxy issues on nodepool. If your behind a proxy, the settings should work. If not, make sure you take out the parts I added. | 21:50 |
| asselin | arosen, otherwise, what I did to debug is use pycharm and set breakpoints. Then you can log into the vm and play around. | 21:51 |
| asselin | arosen, delete these files so that you use o.o upstream's https://github.com/rasselin/os-ext-testing/tree/master/puppet/modules/os_ext_testing/files/nodepool | 21:52 |
| fungi | <public_service_announcement> http://www.openstack.org/vote-paris </public_service_announcement> | 21:53 |
| asselin | arosen, use this line to use o.o upstream (works when there's no proxy) https://github.com/rasselin/os-ext-testing/blob/master/puppet/modules/os_ext_testing/templates/nodepool/nodepool.yaml.erb#L38 | 21:53 |
| *** rainya has quit IRC | 21:53 | |
| arosen | okay i got the command that's failing: http://paste.openstack.org/show/89256/ | 21:54 |
| *** SumitNaiksatam has joined #openstack-infra | 21:56 | |
| arosen | ah okay I see what's causing it. In my env if i do env | grep -i proxy | 21:56 |
| arosen | it has a http_proxy= | 21:56 |
| arosen | once i do unset http_proxy it works fine. | 21:56 |
| arosen | i guess puppet has some problem because of that :( | 21:56 |
| arosen | brb | 21:56 |
| fungi | yeah, looks like the puppet module subcommand literally interprets those proxy envvars and vomits if it can't use their contents as proxy urls. empty variable != missing variable i'm afraid | 21:57 |
| arosen | fungi: can i change base-image to a snapshot with these env variables removed? | 21:59 |
| asselin | arosen, that comes from here: https://github.com/rasselin/os-ext-testing/blob/master/puppet/modules/os_ext_testing/manifests/master.pp#L372 | 21:59 |
| asselin | arosen, https://github.com/rasselin/os-ext-testing/blob/master/puppet/modules/os_ext_testing/templates/nodepool/set_http_proxy_env.sh.erb | 21:59 |
| arosen | Oh so it looks like i can just fix the code without changing the image ? :) | 21:59 |
| asselin | arosen, nodepool only looks at env variables prefixed with NODEPOOL_ | 22:00 |
| *** jp_at_hp has joined #openstack-infra | 22:00 | |
| arosen | asselin: i see in my env I have: | 22:00 |
| arosen | root@Ubuntu1204Server:/home/nicira/os-ext-testing-data# env | grep -i NODE | 22:00 |
| arosen | NODEPOOL_HTTPS_PROXY= | 22:00 |
| arosen | How do i remove those? | 22:01 |
| asselin | arosen, I should probably add conditionals so that they're only exported when non-blank | 22:01 |
| arosen | asselin: can i just comment everything out in: ./os-ext-testing/puppet/modules/os_ext_testing/templates/nodepool/set_http_proxy_env.sh.erb | 22:01 |
| arosen | Then i i restart node-pool it won't have that set? | 22:01 |
| asselin | arosen, yes | 22:01 |
| arosen | asselin: directly from ~/os-ext-testing/puppet this change can be made like this? | 22:02 |
| asselin | arosen, not sure I understand the question | 22:03 |
| *** hogepodge has quit IRC | 22:04 | |
| *** datsun180b has quit IRC | 22:04 | |
| arosen | asselin: i don't think the code really runs out of that dir on my system | 22:04 |
| *** bauzas has joined #openstack-infra | 22:04 | |
| arosen | I think I see these settings also here /etc/profile.d/set_http_proxy_env.sh | 22:04 |
| asselin | arosen, lines 4-6 need to be commented. You need line 7 | 22:05 |
| *** dkliban has quit IRC | 22:05 | |
| asselin | otherwise after you create the node, jenkins won't be able to log in | 22:05 |
| *** pdmars has quit IRC | 22:07 | |
| arosen | asselin: hrm do i need to restart zuul? When i go into my instance i still see this env set. | 22:07 |
| mattoliverau | Morning | 22:08 |
| *** bknudson has quit IRC | 22:08 | |
| asselin | arosen, you need to restart nodepool in a new shell window / clean env | 22:09 |
| *** achanda has joined #openstack-infra | 22:09 | |
| asselin | arosen, no need to restart zuul for this | 22:09 |
| asselin | sudo su - nodepool | 22:10 |
| asselin | set | grep proxy should be clean | 22:10 |
| arosen | cool, lets cross our fingers :) | 22:10 |
| asselin | arosen, back in a bit | 22:11 |
| arosen | sounds good. Thanks for your help! | 22:11 |
| *** maxbit has quit IRC | 22:12 | |
| *** banix has quit IRC | 22:12 | |
| *** e0ne has quit IRC | 22:12 | |
| *** dimtruck is now known as zz_dimtruck | 22:13 | |
| *** flaper87 is now known as flaper87|afk | 22:15 | |
| arosen | asselin: :( when i ssh into the instance it still has http_proxy set :/ | 22:16 |
| arosen | env | grep -i http has nothing for where i'm running nodepoold | 22:17 |
| *** lukego has quit IRC | 22:18 | |
| *** shashankhegde has quit IRC | 22:18 | |
| *** rainya has joined #openstack-infra | 22:18 | |
| *** armax has quit IRC | 22:19 | |
| arosen | looks like something is bringing over an /etc/enviroment file | 22:20 |
| *** derekh__ has joined #openstack-infra | 22:20 | |
| *** dmsimard is now known as dmsimard_away | 22:20 | |
| *** mfink_ has quit IRC | 22:20 | |
| *** derekh_ has quit IRC | 22:21 | |
| *** derekh__ has quit IRC | 22:22 | |
| arosen | asselin: found it. It's happening in: https://github.com/rasselin/os-ext-testing/blob/master/puppet/modules/os_ext_testing/templates/nodepool/set_http_proxy_env.sh.erb | 22:25 |
| *** blkperl_ is now known as blkperl | 22:26 | |
| fungi | fun... marconi is probably going to be renamed to zaqar? that'll be one to keep an eye out for adding to our renames schedule | 22:26 |
| *** kgiusti has joined #openstack-infra | 22:27 | |
| dtroyer | I initally read that as Zarquon… | 22:27 |
| jeblair | fungi: i don't know why we bother naming anything other than "temporary_project_1" | 22:27 |
| clarkb | jeblair: ++ | 22:27 |
| clarkb | I suggested we use uuids at one point | 22:28 |
| fungi | jeblair: clarkb: yes, i vote we name every project `pwgen -s 16 1` | 22:28 |
| fungi | or uuids, why not | 22:28 |
| arosen | asselin: sorry i mean: os-ext-testing/puppet/modules/os_ext_testing/files/nodepool/prepare_node_devstack_http_proxy.sh | 22:28 |
| *** samy_ has quit IRC | 22:28 | |
| *** dims has quit IRC | 22:28 | |
| fungi | dtroyer: at least it wasn't prostetnic-vogon-jeltz | 22:29 |
| *** cnesa8 has joined #openstack-infra | 22:29 | |
| dtroyer | I'm thinking PKI names would be good, then we can encode the repo URLs inside…that sounds good and Vogon-ish | 22:30 |
| dtroyer | and in a year, switch back to UUID names | 22:30 |
| clarkb | jogo: I really really don't want to revert the libvirt-python dep | 22:30 |
| jeblair | dtroyer wins | 22:30 |
| *** mdenny has quit IRC | 22:32 | |
| *** mdenny has joined #openstack-infra | 22:32 | |
| *** cnesa7 has quit IRC | 22:33 | |
| *** garyh_ has joined #openstack-infra | 22:34 | |
| *** shashankhegde has joined #openstack-infra | 22:34 | |
| mordred | clarkb: who is thinking about reverting libvirt-python ? | 22:34 |
| clarkb | mordred: jogo | 22:34 |
| mordred | jogo: why would you want to do that? | 22:35 |
| asselin | arosen, don't use that file. since you're not behind an http proxy, use the o.o file: prepare_node_devstack.sh here: https://github.com/rasselin/os-ext-testing/blob/master/puppet/modules/os_ext_testing/templates/nodepool/nodepool.yaml.erb#L38 | 22:35 |
| *** mdenny has quit IRC | 22:35 | |
| *** mdenny has joined #openstack-infra | 22:35 | |
| fungi | mordred: see ml thread starting at http://lists.openstack.org/pipermail/openstack-dev/2014-July/041509.html | 22:35 |
| *** jgrimm has quit IRC | 22:35 | |
| clarkb | at some point we haev to push into the future and someone will always be burned | 22:36 |
| clarkb | we haven't done this at a fast pace and I think it is completely reasonable as implemented | 22:36 |
| clarkb | it only affects people installing test-requirements.txt | 22:36 |
| mordred | oh - I saw that | 22:37 |
| *** garyh has quit IRC | 22:37 | |
| *** garyh_ is now known as garyh | 22:37 | |
| fungi | right, i think the deprecation cycle argument isn't entirely valid for testing, only for deployment | 22:37 |
| mordred | so install a new libvirt in your deployment | 22:37 |
| mordred | it fits our policy | 22:37 |
| mordred | libvirt is TOTALLY backportable to precise | 22:37 |
| mordred | and having openstack stay on precise libvirt for the lifecycle of precise is crazy | 22:37 |
| clarkb | +1000 | 22:38 |
| clarkb | anyways fungi and I haev responded | 22:38 |
| arosen | asselin: awesome no more proxy env variable. | 22:38 |
| *** elambert has joined #openstack-infra | 22:38 | |
| clarkb | maybe I should've gone to intel today afterall | 22:38 |
| clarkb | but it was too prison like | 22:38 |
| fungi | corporate offices always are | 22:39 |
| sweston | fungi: never got around to the zuul dependency verification yesterday. am starting it now. Shall I assume a start case of a vm with precise installed, no packages? | 22:39 |
| arosen | asselin: I'm guessing now once when the nodepool finishes I should see this node show up in jenkins? | 22:39 |
| asselin | arosen, yes | 22:39 |
| clarkb | fungi: we needed bathroom trip escorts | 22:39 |
| clarkb | so I noped out of there after MAX decided to break yesterday | 22:39 |
| *** signed8bit has quit IRC | 22:40 | |
| fungi | sweston: yeah, we got as far as determining that the zuul python module isn't installing system packages for libxml2-dev and libxslt1-dev i think, which was causing pip install of the zuul source repo to die when python-keystoneclient tried to pull in lxml | 22:40 |
| jhesketh | Morning | 22:40 |
| clarkb | jhesketh: hey tehre | 22:41 |
| sweston | fungi: ok, thanks | 22:41 |
| fungi | sweston: since installing lxml tries to compile system library bindings from scratch on the spot | 22:41 |
| *** Shrews has quit IRC | 22:41 | |
| *** garyh has left #openstack-infra | 22:41 | |
| sweston | fungi: yup. I foresee some weird dependency issues here | 22:41 |
| fungi | sweston: i believe it to be an oversight we introduced when we added python-keystoneclient as a dependency for authenticating swift credential generation | 22:41 |
| clarkb | mordred: fungi don't forget to -1 the revert changes | 22:42 |
| *** david-lyle has quit IRC | 22:42 | |
| fungi | so many things to -1 today... ;) | 22:42 |
| *** david-lyle has joined #openstack-infra | 22:42 | |
| sweston | fungi: ah. ok, good to know. | 22:42 |
| mordred | clarkb: link? | 22:42 |
| fungi | sweston: basically something we've added to zuul since the last time we installed zuul from scratch on a new server | 22:43 |
| clarkb | mordred: https://review.openstack.org/#/c/110773 and https://review.openstack.org/#/c/110774 | 22:43 |
| sweston | fungi: so you are running zuul on baremetal? | 22:43 |
| fungi | sweston: nope, on an ubuntu precise vm in rackspace | 22:44 |
| sweston | fungi: ok | 22:44 |
| mordred | clarkb: -1'd | 22:44 |
| *** prad has quit IRC | 22:45 | |
| *** eharney has quit IRC | 22:45 | |
| *** kgiusti has left #openstack-infra | 22:45 | |
| *** andreykurilin has quit IRC | 22:45 | |
| *** yamahata has quit IRC | 22:46 | |
| *** gondoi is now known as zz_gondoi | 22:46 | |
| fungi | sweston: so i think it should be a relatively trivial patch to add those two system packages via openstack-infra/config:modules/zuul/manifests/init.pp but i suspect nobody has proposed it yet | 22:47 |
| *** david-lyle has quit IRC | 22:47 | |
| *** garyh_ has joined #openstack-infra | 22:47 | |
| *** garyh_ has quit IRC | 22:47 | |
| *** MarkAtwood has joined #openstack-infra | 22:47 | |
| mordred | sweston: we have no bare metal on which to run things :) | 22:47 |
| clarkb | fungi: mordred interestingly removing libvirt-python appears to break unittests? | 22:47 |
| fungi | sweston: also i was waiting for jhesketh and clarkb to possibly chime in on how we managed to get more recent zuul source installed successfully on our server without those system packages in place (since that could point out where/why i'm completely wrong about it) | 22:48 |
| clarkb | or maybe it was the second change. in any case jenkins may prevent merging for us :) | 22:48 |
| *** zns has joined #openstack-infra | 22:48 | |
| clarkb | fungi: bodepd brought it up too and I did not know | 22:48 |
| clarkb | fungi: I think possibly via transitive deps? | 22:48 |
| clarkb | but they are both -dev pkgs so not likely | 22:49 |
| mordred | well, we don't need them any more | 22:49 |
| fungi | clarkb: well, except we have keystoneclient installed without lxml | 22:49 |
| mordred | so it's possible there were there and are not there anymore | 22:49 |
| sweston | fungi: yes, I think you are probably right. However, I need to verify that this is the case before proposing the patch | 22:49 |
| clarkb | fungi: it was pip installed | 22:50 |
| clarkb | then backfilled via puppet iirc | 22:50 |
| clarkb | it just worked (tm) | 22:50 |
| sweston | fungi: is that possible? I thought keystone required prettyxml which required lxml | 22:50 |
| fungi | clarkb: very weird... we have python-keystoneclient==0.7.0 according to pip freeze but no lxml | 22:50 |
| clarkb | fungi: where is the marconi rename discussion? | 22:50 |
| clarkb | fungi: oh you know what | 22:50 |
| clarkb | fungi: wheels maybe? | 22:50 |
| sweston | fungi: apparently it is | 22:50 |
| *** Sukhdev has joined #openstack-infra | 22:50 | |
| clarkb | except we shouldn't publish wheels like that | 22:51 |
| clarkb | but maybe we do | 22:51 |
| *** thedodd has quit IRC | 22:51 | |
| jhesketh | fungi: which system packages are those? | 22:51 |
| fungi | jhesketh: pip packages | 22:51 |
| * sweston hates dependencies on dependencies | 22:51 | |
| *** hogepodge has joined #openstack-infra | 22:51 | |
| fungi | jhesketh: oh, you mean which system packages were preventing compilation of lxml? | 22:51 |
| jhesketh | oh lxml is still being problematic? | 22:52 |
| fungi | jhesketh: libxml2-dev and libxslt1-dev i believe | 22:52 |
| *** james_li has quit IRC | 22:52 | |
| *** garyh has joined #openstack-infra | 22:52 | |
| fungi | are the corresponding system packages that lxml needs to be able to compile its bindings | 22:52 |
| jhesketh | I'm not sure what we did | 22:52 |
| jhesketh | is it possible it's only recently stopped compiling? | 22:52 |
| fungi | it's very mysterious | 22:52 |
| fungi | mmm, doubtful... lxml has needed those since time immemorial | 22:53 |
| fungi | oh, you know what? | 22:53 |
| fungi | it very well may be broken on zuul.o.o too, but pip is getting far enough to install things zuul and keystoneclient actually need without successfully installing things they think they might need (e.g. lxml) | 22:54 |
| mordred | fungi: ++ | 22:54 |
| fungi | so i bet that's why pip freeze says lxml isn't there | 22:54 |
| *** nelsnelson has joined #openstack-infra | 22:54 | |
| fungi | and if we fix the puppet manifest to add the system deps for it, then on the next pip install of zuul it'll complete | 22:54 |
| fungi | i'll check the logs | 22:54 |
| sweston | fungi: shall I continue? | 22:55 |
| *** nelsnelson has quit IRC | 22:55 | |
| *** armax has joined #openstack-infra | 22:55 | |
| jeblair | sdague: can you look at 106460 please? | 22:55 |
| *** nelsnelson has joined #openstack-infra | 22:55 | |
| clarkb | we should just kill xml apis | 22:56 |
| fungi | sweston: i might have a patch for you to test here in a moment | 22:56 |
| sweston | fungi: ok | 22:56 |
| openstackgerrit | A change was merged to openstack-infra/config: Add requests-mock to stackforge https://review.openstack.org/105941 | 22:56 |
| clarkb | fungi: that could be it. especially since no one uses xml apparently | 22:57 |
| *** wenlock has quit IRC | 22:57 | |
| mordred | clarkb: ++ | 22:58 |
| *** jaypipes has quit IRC | 22:58 | |
| *** wenlock has joined #openstack-infra | 22:59 | |
| fungi | huh... so i think i'm not going to find it in zuul' | 23:00 |
| fungi | er | 23:00 |
| fungi | in zuul's puppet logs | 23:00 |
| *** svarnau has quit IRC | 23:01 | |
| fungi | syslog doesn't go back far enough, and all zuul's direct dependencies are already satisfied so pip isn't retrying to update indirect deps | 23:01 |
| sweston | fungi: ah | 23:01 |
| arosen | asselin: hrm after i made that config change it seems the the output from the vm isn't being streamed to nodepoold | 23:02 |
| sweston | fungi: ok, moving on to next plan | 23:02 |
| arosen | but if i ssh to the instance i can see it's doing things. | 23:02 |
| sweston | fungi: how to reproduce in a fresh vm | 23:02 |
| asselin | arosen, not sure...it should still. | 23:02 |
| fungi | seems i would need the puppet apply logs from march when https://review.openstack.org/68297 merged to confirm it in production | 23:03 |
| fungi | which is when we added keystoneclient and swiftclient to zuul's reqs | 23:03 |
| asselin | arosen, instead, stop the nodepoold process and run nodepool image-update <name> <provider> | 23:03 |
| openstackgerrit | A change was merged to openstack-infra/config: Add experimental job Neutron DVR testing on single node https://review.openstack.org/106902 | 23:03 |
| arosen | what's name and provide? | 23:03 |
| *** achanda has quit IRC | 23:03 | |
| arosen | asselin: sorry why do i want to run nodepool image-update ? | 23:04 |
| asselin | arosen, probably d-p-c and local_01 (they're in the /etc/nodepool/nodepool.yaml file) | 23:04 |
| asselin | arosen, you can see the log output better there than when run as part of the nodepoold process | 23:04 |
| fungi | arosen: image-update will do an on-demand build of the image corresponding to the label and provider name you specify, and the results of that session will be streamed to your tty | 23:05 |
| arosen | the image is still in building state though | 23:05 |
| *** thedodd has joined #openstack-infra | 23:05 | |
| arosen | if i do: du -hs in /opt i see it's still downloading stuff. | 23:05 |
| fungi | arosen: is nodepoold logging the image build details to an image.log instead? | 23:05 |
| arosen | fungi: nah still running it in the fg | 23:06 |
| openstackgerrit | Elizabeth K. Joseph proposed a change to openstack-infra/devstack-gate: Updates to README.rst https://review.openstack.org/107427 | 23:06 |
| arosen | i'll give it more time to keep building. | 23:07 |
| arosen | So once i get a job working in jenkins it's going to pick a node that connected to it | 23:07 |
| arosen | then when it's done who deletes the server? | 23:07 |
| fungi | nodepoold does | 23:08 |
| asselin | arosen, that's done here: https://github.com/rasselin/os-ext-testing-data/blob/master/etc/zuul/layout.yaml#L54 | 23:08 |
| fungi | assuming it's successfully connecting to the zmq event stream from your zmq event publisher plugin on the jenkins master | 23:09 |
| asselin | arosen, yes, I didn't update the instructions to enable the zmq event publisher. it should be installed in your system. | 23:09 |
| fungi | that will tell it when to consider the node in-use, and when it's ready for deletion | 23:10 |
| asselin | arosen, the port set for the plugin in jenkins needs to be the same as here: https://github.com/rasselin/os-ext-testing/blob/master/puppet/modules/os_ext_testing/templates/nodepool/nodepool.yaml.erb#L11 | 23:10 |
| jeblair | (todo: move that into zuul) | 23:10 |
| arosen | asselin: so i probably need to update that to the ip of the jenkins server? | 23:11 |
| arosen | since one of the nodepool instances is going to try and connect to localhost:8888 right now? | 23:11 |
| asselin | arosen, it should be localhost | 23:12 |
| asselin | unless the nodepool process runs on a different machine than jenkins | 23:12 |
| *** yamamoto_ has quit IRC | 23:12 | |
| arosen | how long does it normal take to build an image? | 23:12 |
| *** yamamoto has joined #openstack-infra | 23:12 | |
| asselin | arosen, 30 minutes? | 23:12 |
| fungi | depends on how much it needs to do to it. anywhere from a few minutes to (for most of ours) an hour | 23:12 |
| fungi | which is why we schedule daily image updates | 23:13 |
| fungi | rather than something much more frequent | 23:13 |
| *** bknudson has joined #openstack-infra | 23:13 | |
| *** dangers is now known as dangers_away | 23:13 | |
| arosen | so after jenkins deletes the instance because of: parameter-function: single_use_node | 23:14 |
| arosen | does nodepool spawn us another instance since it detected we have one less instance running now? | 23:14 |
| fungi | arosen: jenkins doesn't actually delete it, just marks it offline | 23:14 |
| *** flaviof is now known as flaviof_zzz | 23:14 | |
| asselin | arosen, jenkins tells nodepool the job is done via zeromq publisher. | 23:15 |
| asselin | arosen, nodepool deletes it and spawns a new one | 23:15 |
| fungi | arosen: and nodepool starts building new instances to fill your min-ready as soon as instances transition from ready to in-use for a running job, as long as you have capacity to spare in your max-servers for that provider | 23:15 |
| fungi | arosen: actually, potentially even sooner since it will also take the waiting jobs in zuul into account to determine whether demand exceeds your min-ready pool | 23:16 |
| arosen | So I have min-ready set to 2 but I guess it builds one image and after that is done it will launch the other node based off the first image? | 23:16 |
| asselin | yes | 23:17 |
| fungi | arosen: yeah, image building happens first, then nodes themselves are launched from the most recent ready snapshot (normally when you're building in production you'll already have a previous working snapshot for it to continue using while the new snapshot is being created | 23:17 |
| *** yamamoto has quit IRC | 23:17 | |
| *** yamamoto has joined #openstack-infra | 23:17 | |
| arosen | I'll also see a few snapshots in glance I guess? | 23:17 |
| fungi | presumably | 23:18 |
| arosen | I guess nodepool deletes the old snapshots too | 23:18 |
| fungi | yep | 23:18 |
| *** rcarrill` has joined #openstack-infra | 23:18 | |
| *** bknudson has quit IRC | 23:18 | |
| arosen | cool stuff. I'm pretty excited to see all this work :) | 23:18 |
| fungi | normally it will delete the third-oldest ready snapshot leaving you with two usable snapshots at any point in time | 23:18 |
| fungi | so that if your latest snapshot turns out to be a dud, you can delete it and all subsequent nodes use the one from the previous day | 23:18 |
| *** salv-orlando has quit IRC | 23:19 | |
| fungi | which hopefully isn't suffering from the same defect ;) | 23:19 |
| *** banix has joined #openstack-infra | 23:19 | |
| *** bknudson has joined #openstack-infra | 23:19 | |
| *** rcarrillocruz has quit IRC | 23:20 | |
| *** rainya has quit IRC | 23:20 | |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/config: Zuul depends indirectly on libxml/libxslt headers https://review.openstack.org/110807 | 23:20 |
| fungi | sweston: ^ that's probably what we want | 23:21 |
| *** rainya has joined #openstack-infra | 23:21 | |
| ianw | fungi: would you be ok signing off on https://review.openstack.org/#/c/110178/ (Add experimental Centos7 job) ; nodepool logs seem to show the centos7 nodes built ok, so i'd like to try it | 23:21 |
| asselin | fungi, I'll try that out | 23:22 |
| *** dmsimard_away is now known as dmsimard | 23:22 | |
| *** masayukig has quit IRC | 23:24 | |
| fungi | ianw: double-checking, but then sure | 23:24 |
| *** cdub has quit IRC | 23:24 | |
| sweston | fungi: cool | 23:25 |
| *** pcm__ has quit IRC | 23:26 | |
| sweston | fungi: verification will take until tomorrow, most likely. I need to bring back up my puppetmaster first | 23:26 |
| sweston | fungi: is that ok? alternatively, I can test without a puppetmaster | 23:27 |
| jeblair | jraim__: i did not see the move of kite into barbican program; i think we should move the repo from stackforge/ to openstack/. sound good? any preferred time? | 23:27 |
| fungi | sweston: it's up to you--i'm in no rush | 23:27 |
| jeblair | jraim__: I've added it to the infra team meeting agenda: https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting#Upcoming_Project_Renames | 23:27 |
| fungi | sweston: sounds like asselin is testing it now as well | 23:28 |
| sweston | fungi: ok. even better, two independent verifications | 23:28 |
| asselin | fungi, sweston, yes running now. I didn't set the log ouput = true though, so hopefully it'll work :) | 23:28 |
| sweston | asselin: cool | 23:29 |
| fungi | ianw: looks like we have centos7 images successfully built in all our rax regions for at least the past couple days (presumably longer) and a coupe ready nodes which have been sitting around all lonely for nearly a day waiting on some action, so here you go | 23:29 |
| openstackgerrit | K Jonathan Harker proposed a change to openstack-infra/jeepyb: Give -d precedence over -v in manage-projects https://review.openstack.org/110808 | 23:29 |
| *** masayukig has joined #openstack-infra | 23:31 | |
| *** mancdaz has quit IRC | 23:31 | |
| *** mdenny has quit IRC | 23:32 | |
| *** mdenny has joined #openstack-infra | 23:32 | |
| jeblair | unreviewed_config_patches -= 30 | 23:34 |
| openstackgerrit | K Jonathan Harker proposed a change to openstack-infra/jeepyb: Add timestamps to logging https://review.openstack.org/110813 | 23:35 |
| *** otherwiseguy has quit IRC | 23:36 | |
| fungi | you are a reviewing machine | 23:37 |
| *** dmsimard is now known as dmsimard_away | 23:37 | |
| ianw | fungi: the changes went in yesterday after the meeting, thanks | 23:38 |
| fungi | ianw: awesome | 23:39 |
| *** wenlock has quit IRC | 23:40 | |
| *** wenlock has joined #openstack-infra | 23:40 | |
| openstackgerrit | A change was merged to openstack-infra/config: Update devstack-gate docs in a world with Nodepool https://review.openstack.org/107327 | 23:40 |
| pleia2 | \o/ | 23:40 |
| *** reed has quit IRC | 23:41 | |
| *** wenlock has quit IRC | 23:47 | |
| openstackgerrit | A change was merged to openstack-infra/config: Fail rally-cli job if at least one test failed https://review.openstack.org/107675 | 23:47 |
| *** atiwari has quit IRC | 23:47 | |
| *** thedodd has quit IRC | 23:48 | |
| asselin | fungi, sweston, looks good: http://paste.openstack.org/show/89276/ | 23:48 |
| *** ihrachyshka has quit IRC | 23:49 | |
| openstackgerrit | A change was merged to openstack-infra/config: Move gate configuration to nova-docker. https://review.openstack.org/107794 | 23:50 |
| *** che-arne|2 has joined #openstack-infra | 23:51 | |
| *** thedodd has joined #openstack-infra | 23:52 | |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/config: Zuul depends indirectly on libxml/libxslt headers https://review.openstack.org/110807 | 23:52 |
| *** che-arne has quit IRC | 23:52 | |
| fungi | jhesketh: ^ attempted to address your concerns | 23:52 |
| fungi | asselin: awesome! thanks for testing | 23:52 |
| *** jamielennox|away is now known as jamielennox | 23:53 | |
| jhesketh | fungi: oh so you did decide to use a system package | 23:54 |
| *** gyee has quit IRC | 23:55 | |
| fungi | jhesketh: well, in the end it was either use system packages of lxml's build-time deps, or use a system package of lxml itself. theey were going to be second+ order dependencies of zuul either way | 23:55 |
| *** yjiang5 has quit IRC | 23:55 | |
| jhesketh | yeah fair point | 23:55 |
| *** morganfainberg_Z is now known as morganfainberg | 23:55 | |
| *** amotoki has quit IRC | 23:56 | |
| fungi | what would be cooler would be to figure out if zuul can work with the packages of python-swiftclient and python-keystoneclient present in ubuntu precise, but i have my doubts | 23:56 |
| fungi | that would at least stop us from declaring second-order deps | 23:56 |
| *** banix has quit IRC | 23:57 | |
| *** thedodd has quit IRC | 23:57 | |
| fungi | oh, i should add an inline comment for python-lxml stating that python-keystoneclient requires it | 23:57 |
| *** bauzas has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!